THE TRANSFORMATION RISK-BENEFIT MODEL OF ARTIFICIAL INTELLIGENCE:BALANCING RISKS AND BENEFITS THROUGH PRACTICAL SOLUTIONS AND USE CASES

International Journal of Artificial Intelligence and Applications (IJAIA), Vol.15, No.2, March 2024
DOI:10.5121/ijaia.2024.15201 1
THE TRANSFORMATION RISK-BENEFIT MODEL OF
ARTIFICIAL INTELLIGENCE:BALANCING RISKS AND
BENEFITS THROUGH PRACTICAL
SOLUTIONS AND USE CASES
Richard Fulton,1
Diane Fulton,2
Nate Hayes3
and Susan Kaplan3
1
Department of Computer Science, Troy University, Troy, Alabama, USA
2
Department of Management, Clayton State University, Morrow, Georgia, USA
3
Modal Technology, Minneapolis, Minnesota, USA
ABSTRACT
This paper summarizes the most cogent advantages and risks associated with Artificial Intelligence from an
in-depth review of the literature. Then the authors synthesize the salient risk-related models currently being
used in AI, technology and business-related scenarios. Next, in view of an updated context of AI along with
theories and models reviewed and expanded constructs, the writers propose a new framework called “The
Transformation Risk-Benefit Model of Artificial Intelligence” to address the increasing fears and levels of
AI risk. Using the model characteristics, the article emphasizes practical and innovative solutions where
benefitsoutweigh risks and three use cases in healthcare, climate change/environment and cyber security to
illustrate unique interplay of principles, dimensions and processes of this powerful AI transformational
model.
KEYWORDS
Artificial Intelligence, Risk-Benefit Models, AI Challenges, AI Advantages, Generative AI
1. INTRODUCTION
While Artificial Intelligence (AI) offers many benefits in improving efficiency, accuracy,
accessibility, pattern-recognition and creating higher-paid highly skilled technology-related jobs,
sustainability and quality of life, it also presents risks including invading privacy, displacing jobs,
creating bias, increasing fraud/deception and weaponizing of AI. As the costs and number of risks
associated with AI rise, the ability both to assess those risks and to engage workers in
implementingcontrols will be the new competitive advantage [1].
This article addresses the risks of AI through the lens of benefit-risk analysis, based on relevant
models and theories. From an analysis of AI advantages and risks, the authors review risk
models to create a new paradigm for transforming AI risks into benefits. A set of principles,
dimensions and processes that work in the AI context are laid out from thorough analysis of
previous work, addressing the limitations and omissions of empirical risk models. Next,
innovative and practical solutions are based on principles, dimensions and processes of risks and
opportunities. Lastly, threeuse cases in healthcare, climate change/environment and cyber security
add depth to the body of knowledge and help leaders visualize how to implement balancing risks
with benefits.

2
1.1.Benefits of Artificial Intelligence
From the industrial revolution, significant development in technical innovation has succeeded in
transforming manual tasks and processes that had been in existence for decades where humans
hadreached the limits of physical capacity. It is no surprise that AI offers this same transformative
potential for the augmentation of human tasks within a wide range of industrial, intellectual and
social applications [2]. If society and organizations managed the risks associated with the
industrialrevolution technologies, it is apparent that addressing the challenges of AI is also within
reach.
Faced-paced breakthroughs in algorithmic machine learning and autonomous decision-making
giveleaders new opportunities to balance risks and benefits[2]. Through data analysis, identifying
patterns and making predictions, AI can help society to solve difficult problems like cybercrimes
and create a better world, impacting education, healthcare, clean water, climate change, poverty
andhunger. By harnessing the power of AI, society can create a worldwide sustainable future [3]. In
fact,from a comprehensive analysis study, AI was found to be an enabler on 134 general targets
(79%) across the areas of Sustainable Development Goals (SDGs) set by the United Nations
aimed at overcoming certain pressing worldwide problems.
In the societal and environmental areas of challenges, AI can enhance the provision of food,
health, water and energy services to the population and create smart and low-carbon cities
encompassing a range of interconnected technologies such as electric autonomous vehicles and
smart appliances thatenable demand response in the electricity sector and smart grids that match
electrical demand to weather conditions [4]. By 2030, experts predict 20 to 50 million new AI-
related jobs will be createdglobally and overall spending on technology could increase by more
than 50 percent by 2030 [5]. Studies have analyzed the societal, economic and organizational
impact of this significant change, showing a changing jobs market that is predicted to focus
humans on more creative and higher thinking skills in AI-support roles that pay more [6].
Generative AI is creating more types of jobs [7]. Upwork Research Institute’s new survey of
1,400 U.S. business leaders confirms the positive job impact of AI. Although there may be
variation in generative AI adoption of technologies such as ChatGPT and Midjourney across
companies, two-thirds of top leaders agree that they will increase hiring due to generative AI
[8].
The benefits of applying AI technologies to big data problems are their analytic insight and
predictive capability [9]. Health related studies, for example, show AI technologies can greatly
support patient health-based diagnosis and predictive capability [10][11][12][13]. Economically,
AIgenerates productivity gains [14]. In the legal and policy arena, the best cybersecurity solutions
have AI and machine learning built into them, which creates dynamic protection to face ever-
improving attacks and fraud threats that use automated scripts[15]. Although generative AI helps
those who wish to radicalize people to create deep fake videos used online in social media, AI
tools can also prevent it from succeeding. In November 2022, Intel released FakeCatcher, a
cloud-based AI tool that it claims can accurately detect fake videos 96% of the time, using up to
72 different detection streams [16].
Retail and service-oriented organizations benefit from AI’s ability to personalize the online
experience through real-time, tailored product recommendations based on customers’ purchase
history, age, gender, geographic location, and other data points which continuously upgrades its
capabilities by analyzing millions of customer-service interactions in real time [17]. With the help
of AI, companies can save time and resources and can create even more personalized experiences
directly communicating with their customers, which will result in enhanced brand loyalty
and lifetime relationships [17]. Item recommendation is instrumental for a content provider to

3
grow its audience and evidence shows that automated recommendations account for 35% of sales
on amazon.com, 50% of initial messages sent on match.com and 80% of streamed hours on
netflix.com[18]. Big Data Analytics (BDA) develops the methodological analysis of large data
structures tackling data volume, velocity, variety, veracity and value issues. BDA combined with
AI has the potential to transform areas of manufacturing, health and business intelligence offering
advanced insights within a predictive context [19][20][21]. Organizations are increasingly
deploying data visualization tools and methods to make sense of their big data structures. Because
the analysis andprocessing of complex heterogeneous data is problematic and human perception
and cognition maybe limited, organizations can extract significant value, understanding and vital
management information from big data via intelligent AI-based visualization tools [22][13][23].
AI benefits are organized into societal (SOC), economic (ECON), ethical (ETH), political/
legal/policy (POL), environmental (ENV), data (DATA), technological/implementation (TECH)
and organizational/managerial (ORG) arenas [13] and examples in each category are shown in
Table 1.
2. RISKS OF ARTIFICIAL INTELLIGENCE
The implementation of AI technologies can present significant challenges or risks for
governmentsand organizations as the scope and depth of potential applications increase and the use
of AI becomes more mainstream. Although AI-based systems are increasingly being leveraged to
provide value to organizations, individuals and society, significant attendant risks have been
identified recently. [24] divided risks into two categories including 1) Risks to fundamental
rights—flaws in the overall design of AI systems or biased data can lead to breaches in fundamental
rights, including free speech,discrimination based on sex, race, religion, disability, age or sexual
orientation, protection of personal data and private life and consumer protection; and 2) Risks to
safety and liability—flaws inthe design of AI technologies may present new safety risks for users
related to the availability and quality of data or other problems stemming from AI and machine
learning.
This research builds upon 8 AI challenge categories proposed in a public sector model, used to
organize the risks: 1) societal; 2) economic; 3) political, legal and policy; 4) environmental;
4) ethical; 5) data; 6) technological and implementation; and 8) organizational and managerial
[13]. Public organization adoption of AI and data science presents numerous known challenges
ranging from employee path dependency on embedded processes and norms, information silos,
and a lack of resources, collaborative culture and technical capacities [25][26][27]. AI systems
have had some harmful side effects on communities, through effects on employment and
inequality [28], privacy and safety—injury, property loss and workplace hazards [14], addictive
behavior [29], fairness, bias and discrimination [30][31], human rights [12] and polarization,
extremism, manipulative practicesand conflict [32][33][34][18].
For example, researchers found race bias [35][36] and gender bias [37] for policing algorithms
andbias against working-class and disadvantaged communities for educational assessments in the
UK during COVID [38]. In addition, an AI algorithm created by Zillow was unable to accurately
capture complex assessments such as estimating home values thus causing layoffs of 2,000
employees and a sell-off of its iBuying division [39]. Industry-friendly hackers also fooled Tesla's
Autopilot AI program into merging into oncoming traffic and took control of the car using a video
game controller [40]. To make matters worse, the high legal bar to prove either a disparate
treatment or disparate impact cause of action under Title VII, coupled with the “black box” nature
of many automated hiring systems makes the detection and redress of bias exceedingly difficult
[30]. In the intersectionof privacy and human rights, AI used for target marketing and customer
service may gather data that may include the user’s private behaviors as playing a certain game,
smoking, watching porn or defaulting on loans [18]. Concerns regarding online personalization

4
range from recommendation algorithms isolating information seekers from differing viewpoints
(filter bubble), radicalizing citizens’ attitude towards controversial issues (polarization) or
enabling malicious content [18]. Generative AI has become widely popular, but its adoption by
businesses comes with a degree of ethical risk [41]. With relatively modest amounts of data and
computing power using generative AI,
Table 1. Salient AI benefits by categories and examples.
CategoriesExamples
SOC Worldwide productivity gains; Workers can choose what tasks they want to do; Societal
increased well-being; Impact world sustainability goals of UN; Analytics and AI based
predictions create better patient healthcare; Improve Customer/Organization Interface;
Tackle world problems; Overcome Cultural barriers; Protect Human rights
ECON Increase higher-skill and higher-paying jobs; Create new job categories; Improve
profitability; Optimize returns on investment
POL Detect deep fakes, propaganda, or spying on users; Fight cybercrimes and fraud; Stop
weaponizing AI for harmful purposes; Add rules and accountability in the use of AI; Rely on
European Union guidelines/rules for responsible AI
ENV Address climate change and resources scarcity; impact carbon footprint.
ETH Build trust towards AI-based decision making; Increase explanation of decisions made by AI;
improve processes relating to AI and human behavior, find balance in machine versushuman
value judgements; be vigilant in testing for bias/discrimination
DATA Advanced insights within a predictive context; improve 5 V’s of data; Increase transparency
and reproducibility; Create reliable/sufficient data pools; Improve data integration and
continuity; Create standards for data collection
TECH Fight adversarial attacks; Increase transparency and interpretability; Smart design of AI
systems; Increase AI safety and security; gain value from big data; continue to innovateways
to decrease problems using unstructured data; integrate legacy and new systems
through collaboration with vendors, governmental and educational institutions
ORG Personalize customer experiences; Streamline repetitive and boring jobs; Visualize to
increase understanding of firm needs; Improve data sharing and collaboration; Increase AI
talent through HR innovation in hiring and retraining; Create an AI-driven culture
the average person can create a video of a world leader confessing to illegal activity leading to a
constitutional crisis, a military leader saying something racially insensitive leading to civil unrest
inan area of military activity, or a corporate titan claiming that their profits are weak leading to
globalstock manipulation. These so called ‘deep fakes’ pose a significant threat to our democracy,
national security and society [32].There are also practical issues over how accurate machine
learning solutions are. The range of testing approaches available within machine learning is
growing rapidly,and that is a good thing, but it is also driven by the evident limitations of the
previous methods andthe need to overcome those limitations [42].
Another crucial factor is the availability of suitable data. Although machine learning packages
for Python and R can easily read all types of data from Excel to SQL and can perform natural
languageprocessing and process images, the speed with which machine learning solutions have
been proposedhas not kept pace with firms’ abilities to suitably organize the internal data they
have access to. Datais often held in separate silos across departments, on different systems and
with internal political and regulatory issues restricting the sharing of data. Important data might
not even be recorded as data but rather kept as informal knowledge of the firm [42]. Estimates for
work displacement due to automation highlight that up to a third of current work activities
(between 400 million and 800 million jobs around the world) could be impacted by 2030 [5]. A

5
survey of the top 1000 firms in the United States found that their biggest concern in the
implementation of AI was the readiness and ability of staff to understand and work with these
novel solutions [43]. Sixty-four percent of US executives and 70 percent of European leaders
believe they will need to retrain, up-skill or replace a fourth of their workers due to advancing
automation and digitization [44]. Please see Table 2 fora summary of salient AI risk examples
by categories.
3. MODELS AND THEORY
Very few comprehensive models of AI risk are available today to help managers assess and deal
withthe increasing risks they face. To address this void and the ever-increasing depth and breadth
of AItechnology challenges, the researchers critically review existing related risk models and start
with an historical empirically-based risk model dealing with the adoption of information
technology projects. Since the risk of IT projects failing was extremely high in those days, [45]
developed a model to predict project risk and delineate risk factors associated with organizational
IT projects, which serves as a basis for a newly proposed AI risk framework.
The McFarlan Risk Model (MRM) provided a useful and measurable approach for the diagnosis
andmitigation of IT project risks with three dimensions based on ‘project size’, ‘project structure’
and ‘experience with technology’. For example, if there were excessive costs, large numbers and
levels of staff needed, increased completion time to complete and impacted many different
departments, then the risk was increased. If the user department needed to change a lot of
procedures and structures to meet the project requirements and users were highly resistant to
changes, the project was also considered higher risk. Lastly, if the team lacked the appropriate
experience with the recent technologies, the firm needed to hire more experts or use outside
consultants and if the experts did not work in partnership with the company, the risks expanded
exponentially [45].
These three dimensions were expanded in this research to include ‘resources’ (not just the size of
a project but how many, how well and how long resources support AI implementation),
‘governance’ including guidelines and structural processes put in place (not just the project
structure) and ‘expertise building’ or on-going capacity and commitment to train, retrain, upskill
and encourage employee development (not just relying on the current experience of the workers or
hiring outsiders). In their “Risk Assessment Framework” on implementing enterprise resource
planning projects, risks from ‘external engagement’, program management, work stream and
work package levels across technical, operational, business and organizational categories were
successfully mitigated using on-going risk controls [46].
The specific content of corporate governance guidelines and policies is an important variable to
anyrisk mitigating model. Using a comparison between Belgium (weaker country guidelines) and
Australia (stronger governance guidelines), researchers found significantly more developed risk
management and internal control systems in Australian companies versus Belgian companies
overall[47]. In addition to strong governance guidelines, impactful ‘governance’ also dictates that
an entirerisk management structure be in place. For example, the structure of risk control may
include 4 phases of identifying, assessing, mitigating and monitoring risks [48] or 4 specific
functions— govern, map, measure and manage — to help organizations address the risks of AI
systems [49].
In this research, the ‘governance’ dimension of risk focuses on a broader range of AI policies,
controls and structure. Using a comprehensive AI risk management system can maximize the
benefits of AI technologies while reducing the likelihood of negative impacts to individuals,
groups, communities, organizations and society” [49]. In their survey of 576 companies and a

6
review of 2,750 company/analyst reports, researchers found that financial performance was
highly correlated with the level of integration and coordination across risk, control and
compliance functions. Effectively harnessing AI technology to support risk-benefit management
is the greatest weakness or opportunity for most organizations [50]. To improve the predictive
ability of McFarlan’s model, the “Extended McFarlan Risk Model” (EMRM) was developed to
differentiate between project success and failure, adding an organizational ‘culture’ dimension
[51]. The ‘culture’ attributes were quantitatively measured including users’ practices, users’
attitudes, company working practices, organizational polices such as information technology
policy and data flow practices, internal and external communication practices in the organization,
openness to change and cross-functional coordination. If the corporate culture was lacking, the
implementation risk increased [51]. In this research, the dimension of ‘culture’ was widened to
encompass an ‘AI-driven culture’ with communication, work policies/practices, change
management, innovation and cross- disciplines/perspectives which fully integrate AI throughout
the firm for triple bottom line goals of profits, people and the planet [52]. The new risk-benefit
model proposed is also based on the useof social cognitive theory which views people as active
agents who can both influence and can be influenced by their environment [53].
Table 2. Salient AI risks by categories and examples.
Categories Examples
SOC Customer/Organization Interface; World crises; Cultural barriers; Human rights; Country
specific data profiles; Unrealistic expectations towards AI technology; Country specific
organizational practices and insufficient knowledge on values/advantages of AI.
ECON Affordability of required computational expenses; High costs for customers; High cost
andreduced profits for organizations; Wider divides in society leading to social upheaval
POL Copyright issues; Embedded bias and discrimination by humans or technology; Injury;
Governance of autonomous intelligence systems; Responsibility and accountability;
reduced privacy/safety; National security threats from foreign-owned
companies/governments; Cybercrimes; Weaponizing AI for harmful purposes; Lack of rules
and loss of accountability; Costly human resources still legally required to account
for AI based decision; Lack of official industry standards of AI use and performance
evaluation; difficult to redress bias legally
ENV Impact on carbon footprint, use of resources such as fossil fuels, water, electricity; impact
on climate change and air quality
ETH Lack of trust towards AI based decision making and unethical use of shared data.
Responsibility and explanation of decision made by AI; processes relating to AI and
human behavior, compatibility of machine versus human value judgement, moral
dilemmas and AI discrimination
DATA Lack of data to validate benefits of AI solutions; Format, quantity and quality of data;
Transparency and reproducibility; Insufficient size of available data pool; Lack of data
integration and continuity; Lack of standards of data collection
TECH Adversarial attacks; Lack of transparency and interpretability; Design of AI systems; AI
safety; Specialization and expertise; Big data; Architecture issues and complexities in
interpreting unstructured data; integrating legacy and new systems
ORG Realism of AI; Better understanding of needs of the organizational system; Organizational
resistance to data sharing and collaboration; Lack of inhouse and interdisciplinary AI
talent; Threat of workforce layoffs and retraining; Lack of strategy for AI development;
Embedded processes and norms; lack of resources; information silos
That is, people learn or are transformed by observing and interacting with others. For example, in
her research, success in using cloud computer services was dependent on factors such as ‘external
engagement’ with people who were using the technology and workers’ personal attitude towards
risk and innovation [54]. Lastly, this paper offers a new dimension of ‘transformational
collaboration’ (not just ‘external engagement’ outside the firm construct from previous research)

7
across boundaries such as disciplines, departments, divisions, firms, industries, societies and
countries which emphasizes that a person’s attitude and collaboration with others who are using
innovative technologies such as AI or those who have divergent backgrounds can help increase
technology implementation success.
Combining a systematic and critical literature analysis with related risk models and theory, this
research adds value to the field of AI by synthesizing and creating an integrated model for
managing risk and benefits designed particularly for AI adoption and implementation. Using [45]
as a theoretical lens, the updated McFarlan model [46] for the context of culture and social
cognition theory’s [53] contributions of personal attitude and engagement from previous risk
research [46], “The Transformation Risk- Benefit Model of Artificial Intelligence” presents an
innovative model that expands to 5 broader dimensions of ‘resources’ (RES), ‘governance’(GOV),
‘expertise building’(EXP), ‘AI driven culture’ (AI CUL) and ‘transformational collaboration’ (TR
COL) that has the potential to change AI risks into benefits as illustrated in Figure 1.
4. MODEL PROCESSES
To fully implement “The Transformation Risk-Benefit Model of Artificial Intelligence”, the
authorsanalyze the processes that increase transformation of risks into benefits and propose the
best fit for the AI context. To complete this assessment, a thorough literature review of risk
processes was conducted. For example, risk processes for construction projects included
identification, ranking, outcomes, monitoring and risk response [55]. To manage enterprise
resource planning (ERP) implementation risks, scientists proposed 7 processes of context
analysis, risk identification, risk analysis, risk evaluation, risk treatment, monitoring and review,
and communication and consulting[56].
= McFarlan’s Portfolio Approach to Information Systems [45].
= Extended McFarlan Risk Model adding Culture [51].
= Risk Assessment using Social Cognitive Theory [46][54].
Figure 1. “The Transformation Risk-Benefit Model of AI”: Transform AI risks into benefits
White
Black
Grey

8
The railway industry uses three risk management phases: 1) risk identification, 2) risk assessment
and 3) applying risk reduction plans [57]. Several actions, including risk mitigation and re-
engineering to control risks, are included in traditional risk management models [58]. A four-step
framework of risk assessment based on ISO standards includes (1) risk identification; (2) risk
analysis (3) risk evaluation; and finally (4) risk response planning [59[60].
In medical research, there are five standard processes of conducting a risk-benefit assessment to
determine the level of institutional review (IRB): 1) assess potential risks and discomforts
associated with each intervention or research procedure; 2) estimate the probability that a given
harm may occurand its severity; 3) explain measures that will be taken to prevent and minimize
potential risks; 4) describe the benefits for subjects; and 5) discuss the potential societal benefits
[61].
The general action and solution-related framework of risk management reflects that “for each
issueor event requiring a decision, managers can benefit from adopting a systematic approach
[60]. The main processes of risk assessment in business continuity planning are the identification of
risks, theiranalysis and evaluation, and risk treatment and response planning [62].The drawbacks
to these models are single-minded focus on risks only and not full consideration of benefits and
opportunitiesAI can provide. After a systematic review, the researchers believe the medical model
comes closest to addressing these limitations. In the same vein, an updated ISO 31000 was
proposed in 2018 and expands by including how to create and protect value or benefits for
stakeholders—a positive focusof AI risk-benefit assessment that was previously lacking [63].The
idea is to try to identify all the consequences of a particular issue or event, to find an optimal
decision set to minimize adverse effects and maximize societal and business objectives in a cost-
efficient manner” [64]. This principle of ‘balancing’ risks and benefits is essential to “The
Transformation Risk-Benefit Model of AI”. Another principle of the new risk-benefit model is to
be ‘integrative’. Risk factors, events, benefits and consequences of AI all should be ‘integrated’ to
meet specific organizational and societal needs.For example, in the context of the supply chain,
risks could emerge from supply chain disruptions,which often create material or financial losses,
delivery delays and decreased reputation and competitiveness. However, organizations that learn
to successfully manage these risks can also create benefits for performance and reputation
[65][66].
For small and medium sized enterprises (SMEs), bankruptcy risk is one of the most important
risks identified and using a simulation model of case companies illustrates the importance of
looking at ‘integrative’ factors. Namely, performance on several financial measures such as
customer accounts receivable flow time and credit limit allowed by financial partners were
predictors of bankruptcy risk which lowered the risk from 80% to 30% [67].Creating “financial
sustainability” by assessing financial and nonfinancial indicators and metrics is the answer to
present-day challenges of a very turbulent environment. The main goal of financial sustainability
analysis is to search, discover and mobilize resources, and uncover opportunities enabling
development and success of the organization [68].This approach highlights the importance of a
‘strategic’ outlook – what is good for the firm and society over the long haul and is another
principle needed to transform risks into benefits in AI.
Suggestions made by the assessors for the improvement of the benefit-risk assessment of medical
research/procedures included 4 processes: 1) A well-structured presentation of the relevant data
on benefits and risks; 2) A common unit to compare benefits and risks; 3) A thinking tool/model
that could add transparency; 4) A tool to help deal with uncertainty and complicated data [69]. A
transformational mindset needs to ‘balance’ delivery discipline and accountability, agility and
pragmatism, continuous improvement and a sense of chronic unease. This delicately balanced
mindset enables organizations to course-correct and to quickly address emerging challenges that

9
can be turned into competitive opportunities [70].Traditional risk models are limited by their
standardized, reactive and sporadic approaches, management of risks one by one, and a myopic
risk-averse mindset [71].
Due to its predilection towards emphasizing negative scenarios - not to mention its limited scope
- many organizations have found traditional risk management to be lacking in its ability to
provide sufficient insights and a shaky foundation on which to make informed decisions [72].
Sorting out tangled risks can only be achieved by proactively understanding and managing the
attributes of interconnected risks and benefits [73].To address the limitations and drawbacks of
current risk models, the authors recommend working on benefits and risks simultaneously, using
five processesto ‘assess’ all risks and opportunities, ‘identify’ the level and priority of risks and
opportunities, ‘analyze’ all the various options, plans, and strategies tied to these risks and benefits,
‘create optimalplans’ through collaboration across inside and outside boundaries and ‘evaluate and
adjust plans’ ona continuous improvement loop.
The authors also add to the body of knowledge and expand their Model aimed at transforming AI
risks to benefits by initiating five overriding principles of being ‘holistic’, ‘integrative’,
‘strategic’, ‘proactive’ and ‘balanced’. ‘Holistic’ means looking at events, both positive and
negative, that might impact the organization and society taking the long term big picture into
account. ‘Integrative’ refers to seeing the synergy between solutions and across
departments/divisions/industries. ‘Strategic’ is to have an opportunity-creating rather than a risk-
adverse mindset to make plans that create outcomes to optimize the varied goals of stakeholders.
A ‘balanced’ approach is to look at both negative and positive AI events, working out solutions
that try to minimize risks, increase opportunities and transform risks into benefits. Lastly, the
‘proactive’ principle in the AI context isto think ahead and have plans that continually look to
decrease risks, anticipate handling negative events while searching for opportunities in AI. For an
overview of the principles, dimensions and processes to balance risks and benefits in “The
Transformation Risk-Benefit Model of Artificial Intelligence”, please see Figure 2.
5. PRACTICAL SOLUTIONS
Corporate commitment should extend beyond corporate interests, to address broader societal
responsibilities and ethical AI practices [41]. Leaders remain in need of conceptual, technical and
institutional mechanisms to assess how to achieve accountability for the harmful consequences of
data-driven algorithmic systems [74]. In this section, the authors give examples of pragmatic and
innovative solutions that can transform risks into benefits in each of the Model’s 5 risk
dimensions. Under the dimension of ‘resources’, innovative and practical solutions such as
developing policies to assure AI will be directed at ‘humanness’ and common good and building
inclusive, decentralizedintelligent digital networks ‘imbued with empathy’ will help leaders meet
social/ethical goals [75].

10
Figure 2. Principles, dimensions and processes to balance risks and benefits in “TheTransformation
Risk-Benefit Model of Artificial Intelligence”
Another resource-related solution is to alter organizational, economic and political systems to
better help humans ‘race with the robots’ and reorganize these systems toward the goal of
expanding humans’ capacities and capabilities [75]. Beyond retraining, a range of resource
policies can help, including unemployment insurance, public assistance in finding work and
portable benefits that follow workers between jobs [5]. Increasing human/AI collaboration and
human interventions in decision-making in risky situations like finance, healthcare and hiring
practices will ensure trust/transparency and decrease harms to people [41]. Under the
‘governance’ dimension, comprehensive AI lifecycle governance where policies and procedures
are described and enforced throughout the AI system lifecycle can increase transparency— AI
factsheets are one practical to capture metadata automatically, making enterprise validation and
external regulation easier to monitor [76]. To increase fair and unbiased ‘governance’, create ‘an
auditing imperative’ for algorithmic hiring systems that mandates regular internal and external
audits of automated systems, based upon precedents of Occupational Safety and Health
Administration (OSHA) audits in labor law or the Sarbanes-Oxley Act audits in securities law to
increase trust and confidence in AI- automated systems [30].
Other effective ‘governance’ solutions are for organizations to conduct their own bias,
explainability and robustness assessments to protect privacy and harmful outputs and security
assessments to help identify and decrease threats such as cybercrimes. Also prioritize the
responsible use of AI and generative AI by using “zero or first party data, keeping data fresh and
well labeled, ensuring there’s a human in the loop, testing and re-testing and keeping feedback
close” [41]. Base ‘governance’ onprotecting society and the firm by following stricter regulations,
guidelines & certification standardssuch as those recommended by the European Union [24].
To increase ‘expertise building’, train or retrain workers in new fields such as digital forensics,
cybercrime specialists, data analysts and AI governance specialists, use Protective Optimization
Technologies (POTs) to both explore the societal effects that algorithms and optimization systems
and design countermeasures to eliminate their negative effects [77]. Widen educational
opportunities to step into new AI jobs and learn new skills (throughout the job and educational
systems available or create new ones), build talent in technical competencies like AI engineering
and enterprise architecture and empower people to work effectively with AI-infused processes
[75]. In the arena of creating an ‘AI driven culture’, use AI-based visualization tools [13][23],

11
leverage open-source and user-provided data, be transparent when AI has created content using
watermarks or in-app messages, cite the sources for creating content, explain why the AI gave the
response it did, highlight uncertainty and creating guardrails, extend the commitment to using AI
tools and technologies beyond immediate corporate interests, help meet broader societal
responsibilities and ethical AI practices and prevent some sensitive, risk heavy tasks from being
fully automated [41].
These practical and innovative solutions for developing an ‘AI driven culture’ will have
important benefits such as extracting significant insights from big data and allowing workers to
understand theimpact of AI directly, increasing transparency, accuracy and honesty about data
uses and sources, mitigating risks and biased outcomes while protecting sensitive data processes
and uses [41]. In thefinal dimension of risk, ‘transformational collaboration’ solution include: 1)
collaborating to minimize the size of AI models [41], 2) working with partners to create better
data and train on models with large amounts of high-quality customer relationship management
data; or 3) increasing collaboration, risk prevention and control across departments, disciplines,
companies, countries andsocieties [75].
Working collaboratively can produce outstanding outcomes while lowering threats such as
reducing the carbon footprint because less computation is required in smaller models, which
means less energy (water and electricity) consumption from data centers and carbon emissions,
higher quality data helps to maximize data accuracy, reliability and value, and working across
boundaries and perspectives can facilitate the innovation of widely accepted approaches aimed at
tackling wicked problems and improves risk control over complex human-digital networks
worldwide [75]. Table 3provides a sampling of practical and creative solutions organized into the
five dimensions of risk.
6. USE CASES FOR AI
In addition to elaborating on how risks can be transformed into benefits through a sampling of
doableinnovative solutions, this research focuses on practical use cases in AI from three different
sectors: healthcare, climate change/environment and cyber security.
6.1. Healthcare: Challenges and Solutions
Use cases in the healthcare sector encompass both depth and breadth in their purview including
virtual care delivery, digital health driven by wearables and online medical devices, AI
automationof administrative tasks, enhanced point-of-care decision systems and therapeutic access
through realworld data. The top 3 challenges for AI in healthcare are 1) trust; 2) selecting the right
use cases and 3) managing data volume along with privacy issues [78]. Through analyzing such
big data implementation cases, scientists will be able to understand how AI technical capabilities
transform organizational practices, thereby generating potential benefits in industries such as
healthcare [79].
One of the most recent use cases of AI’s impact on healthcare is its use to fight diseases such as
theCOVID virus. Some generative, deep learning and machine learning methods have been used to
help fight COVID-19. The main advantage of AI-based platforms is to accelerate the process of
detection, diagnosis and treatment of epidemics. These AI platforms work with various data
types, such as clinical data and medical imaging which improve AI approaches to find the best
responses [80].
Challenges still exist, including limited data accessibility, the need for external evaluation of AI
models, the lack of awareness of AI experts of the regulatory landscape of AI tools in healthcare,

12
theneed for clinicians and other experts to work with AI experts in a multidisciplinary context and
the need to address public concerns over data collection, privacy, and protection. Having a
dedicated team with expertise in medical data collection, privacy, access and sharing, AI
scientists handing over training algorithms to the healthcare institutions to train models locally,
and taking full advantage of biomedical data can alleviate some of problems posed by these
challenges and will mean AI can help create practical and useful solutions for combating
pandemics [81]. Another example where AI technology is evolving in the medical/healthcare
sector is using AI big data for the diagnosis, treatment and prognosis of numerous neurological
diseases. Huge amounts of data are gathered from various digital sources, including electronic
health records, media and databases, thenanalyzed, and AI algorithms are trained to detect risk
factors, diagnose diseases and suggest best therapies [82]. It is estimated that AI applications can
cut annual healthcare costs in the U.S. by $150billion in 2026 by changing the healthcare model
from a ‘reactive’ to a ‘proactive’ approach, focusing on health management rather than disease
treatment, resulting in fewer hospitalizations, less doctor visits and less treatments [83].
AI has the potential to inspire systematic ways to process clinical and molecular information that
span the 5Vs challenges of value, volume, velocity, variety and veracity, while identifying
patternsof care, analyzing unstructured data, supporting decisions, prediction, and traceability of
healthcare data [84].The main trend in the healthcare industry is a shift in data type from
structure-based to semi-structured based such as home monitoring, telehealth, sensor-based
wireless devices and unstructured data including transcribed notes, images and videos. The
increasing use of sensors andremote monitors is a key factor supporting home healthcare services
but the challenge is to managethe growing data generated to grow significantly and improve the
quality of healthcare [85]. The challenge posed by clinical data processing involves the quantity
and volume of data but also the varied types of data. AI has the potential in healthcare
organizations to support clinical decision- making, disease surveillance and public health
management [86]. Poor medical data processing systems are the key reasons for medical errors but
using AI data management systems reduces errorsand associated harms increasing patient safety
and efficient data management [87]. Analyzing findings from previous clinical studies could
translate new research conclusions into routine clinical processes and thus drive successful
evidence-based medicine [88]. For example, Optum Labs working with Mayo Clinic, combined
diverse collaborator perspectives with rich data, including deep patient and provider information,
to reveal new insights about diseases, treatments and patients'behavior. Practitioners' involvement
in agenda setting and continuous feedback loops improved AI-driven processes and dimensions
[89]. The Rizzoli Orthopedic Institute in Bologna, Italy analyzed patients’ genomic data and case
histories to determine hereditary diseases risks and to provide information of effective hereditary
disease treatments and develop more evidence-based surgery protocols for genetic disease
patients, reducing imaging requests by 60% [90]. Being able to input clinical information into
visual dashboards, increase real-time information monitoring with alerts and notifications, and
real time data navigation helps healthcare leaders to make sound operational decisions [91]. With
real time dashboards to monitor patients’ health and prevent medical accidents, two case
hospitals—Mental Health Center of Denver and Kaiser Permanente Northern California— were
able to improve quality and addressed medical errors, various patient safety issues and
appropriate medication use [92]. Adapting to AI-related challenges and providing constructive
solutions will require a multidisciplinary approach, innovative data annotation methods and the
development of more rigorous AI techniques and models. By ensuring cooperation between AI
andhealthcare providers and merging best practices for ethical inclusivity, software development,
implementation science, and human-computer interaction, the AI community can create an
‘integrated’ and ‘holistic’ best practice framework [93]. Additionally, collaboration between
multiplehealth care settings to share data, ensure its quality, and verify analyzed outcomes are
critical to success [94].

13
Table 3. Practical and innovative sample solutions by risk dimensions transforming
AI risks intobenefits
DimensionsPractical and Innovative Sample Solutions Transforming AI Risks into Benefits
RES
 Prioritize people by reorganizing
organizational, economic and political systems
towards the goal of expanding humans’
capacities and capabilities
 Increase human/AI collaboration and human
interventions in high risk situations.
 Helps humans ‘race with the robots’ and
staunches AI trends that would
compromise human relevance
 Improves trust and removes harms in
finance, healthcare or hiring.
GOV
 Build comprehensive AI lifecycle governance
where policies and procedures are described
and enforced during the design, development,
deployment, and monitoring of an AI system.
 Create AI factsheets to capture AI metadata
across the model lifecycle automatically,
 Create “an auditing imperative” mandate for
both internal and external audits of automated
hiring systems
 Keep and audit records of job applications.
 Use AI tools for security and threatassessments
 Conduct bias, explainability and robustness
assessments
 Base governance on protecting society from
harm and unwanted negative consequences for
the firm by following stricter regulations,
guidelines & certification standards
recommended by the European Union
 Test and re-test and get feedback close tosource
 Builds trust in AI through transparency
 Facilitates enterprise validation or
external regulation
 Builds trust and confidence in decision
making and has been used in other areas
of law, such as Occupational Safety and
Health Administration (OSHA) audits in
labor law or the Sarbanes-Oxley Act
audit requirements in securities law
 Helps organizations identify
vulnerabilities that may be exploited by
bad actors such as cybercrimes
 Protects privacy and harmful outputs
 Self-regulation decreases risk of harms
 Ensure data is high quality, accurate and
has value
EXP
 Train or retrain workers in new fields such as
digital forensics, cybercrime specialists, data
analysts and AI governance specialists
 Protective Optimization Technologies (POTs)
systematizes the use of technologies as tools to
both explore the effects that algorithms and
optimization systems have on our society, and
the design of countermeasures to contest their
negative effects
 Widen educational opportunities to step into
new AI jobs and learn new skills
 Building talent in technical competencies like
AI engineering and enterprise architecture and
training people across the organization to work
effectively with AI-infused processes.
 New jobs, tools and skills can help
detect deep fakes, harmful patterns inthe
digital arena and make workers more
productive
 Brings structure and strategies to AI tool
and skill use that help not harm society
 More educational options along with
funding (grants, scholarships,
internships) will increase the availability
of skills throughout the organization
andsociety
 Benefits the organization by increasing
efficiency and gives workers higher
skilled higher paid jobs
AI CUL
 Use AI-based visualization tools
 Leverage open-source and user-provided data.
 When autonomously delivering outputs, be
transparent that an AI has created the content
using watermarks on the content or through in-
app messaging
 Responsibly use AI in maintaining accuracy,
safety, honesty, empowerment and
sustainability by citing the sources from where
the model is creating content, explaining why
 Extracts significant value and key
management information from big data
 Ensures honesty about the use and
sources of data
 Increases transparency
 Mitigates risks and eliminates biased
outcomes.
 Improves the accuracy of data by being
transparent and in the organization and
broadly across societies.

14
the AI gave the response it did, highlighting
uncertainty and creating guardrails
 Extend commitment beyond immediate
corporate interests, encompassing broader
societal responsibilities and ethical AI
practices
 Prevent some tasks from being fully
automated
 Human intervention may protect certain
risky & sensitive data processes and
uses.
TRAN COL
 Work collaboratively to minimize the size of
AI models
 Collaborate with vendors and partners to
create better data and train on models with
substantial amounts of high-quality customer
relationship management or other generated
data.
 Collaborate across departments, disciplines,
companies, countries and societies to solve
bigcomplicated or wicked problems.
 Reduces carbon footprint because less
computation is required, which means
less energy consumption from data
centers and carbon emissions
 Maximizes data accuracy, reliability and
value
 Facilitates the innovation of widely
accepted approaches aimed at tackling
wicked problems and maintaining
control over complex human-digital
networks worldwide
6.2. Climate change/Environment: Challenges and Solutions
Leveraging the opportunities offered by AI for global climate change while also limiting its risks
is a path that requires responsive, evidence-based and effective governance to become a
competitive strategy. AI tools are transforming data-driven science — better ethical standards and
more robust data curation are needed [95]. Advances in AI are helping to make sense of all the
copious amounts of data and through training, machine-learning methods get better at finding
patterns without being explicitly programmed to do so [96]. In environmental sciences,
technologies ranging from sensorsto satellites are providing detailed views of the planet, its life
and its history, at all scales with AI applications for weather forecasting [97] and climate
modelling [98] and for assessing damage during disasters to speed up aid responses and
reconstruction efforts [95].
‘Collaboration’ is essential for impacting climate change and the environment since it reduces the
chance of failure such as working on a problem that is not actually impactful, overly simplifying
a complicated issue or using advanced computational tools when simple tools are adequate [99].
An exemplary use case is an open-source project that was sponsored and developed by the
Pentagon’s Defense Innovation Unit and Carnegie Mellon University's Software Engineering
Institute in 2019 called xView2, which uses collaboration with many research partners.
Combining machine-learning algorithms with satellite imagery identifies building and
infrastructure damage in a disaster area and categorizes severity levels much faster than with
current methods [100].
Specifically, this technology was used in California and Australia for wildfire responses [101],
during recovery efforts after flooding in Nepal [102], and in Turkey during several earthquakes to
successfully help workers be “able to find areas that were damaged that they were unaware of,” in
collaboration with Turkey’s Disaster and Emergency Management, the World Bank, the
International Federation of the Red Cross and the United Nations [103]. Although the AI
technologyis quite amazing, it too has problems such as its dependency on satellite images which
are only clearduring the day with no cloud cover and when a satellite is overhead, in remote areas
there are far fewer satellite images and it does not work on the side of a building since an aerial
perspective is needed. Some personnel may not trust the AI technology. New imaging called
synthetic aperture radar, which creates images using microwave pulses rather than light waves,

15
could improve on satellite technology and training of personnel to use and believe in the
technology [100].
Enhancing energy efficiency can significantly contribute to reducing the impacts of climate
change/environment [99]. AI applications such as smart manufacturing can reduce energy
consumption, waste and carbon emissions by 30–50% and energy consumption in buildings by
30– 50%. Nearly 70% of the global natural gas industry applies AI technologies to increase
weather forecasting accuracy and reliability. Combining smart grids with AI optimizes power
systems’ efficiency, to reduce electricity bills by 10–20%. Intelligent transportation systems can
reduce carbon dioxide emissions by approximately 60%. Management of natural resources and
designing resilient cities using AI can further promote sustainability [104].
6.3. Cyber security: Challenges and Solutions
Cybercrime is one of the most prominent domains where AI has begun demonstrating valuable
inputs and is being deployed as the first line of defense. Because AI can detect new assaults faster
than humans, it provides better protection against cybercrime [105]. Cyber security threat
modeling with appropriate Intelligent User Interface (IUI) design helps developers enhance
flexibility, usability and interaction to improving computer-human communications can reduce
critical design time and create a better risk-benefit model [106]. Cyber security systems protect
assets, devices, data, programs and networks from digital attacks which prevent unauthorized
access [107].
Another security concept of threat modeling attempts to enhance cyber security by analyzing,
itemizing and prioritizing the potential threats using the attacker’s point of view [108] prioritizes
strategies against an attack. These steps mirror the Model’s processes of ‘assessing’,
‘identifying’, ‘analyzing options’ and ‘creating optimal plans’. Neglecting the security designers in
cyber securityand focusing only on human factors in cyber-attacks [109] allows wide penetration
to sensitive systems through social engineering and cyber-attacks that militarizes information
technology [110]. Cyber security cuts across technical and social matters and needs to be
prioritized in the movement of businesses and government activities to online platforms have
changed [107]. Therefore, the useof Application Program Interface without the system engineer’s
control undermines the system properties and is an excellent example of the importance of being
‘strategic’ and ‘integrative’ [111].
Leaders also need to be aware of the creation of different laws and regulations with various
provisions for cyber security data protection and privacy that are leading the way around the
world.First is The General Data Protection Regulation (GDPR), a groundbreaking data privacy
regulationcreated by the European Union in 2018, which revolutionized the way personal data is
handled, processed and safeguarded, not only within the EU but globally. The California
Consumer Privacy Act (CCPA) is another landmark privacy legislation enacted in 2020, which
improves on non- discrimination, individual rights and transparency principles being used not
only in California but also across the U.S. Another law includes The Cybersecurity Law of the
People’s Republic of China, a comprehensive legislation enacted in 2017 which safeguards
cyberspace, protects individual/organizational rights and strengthens cybersecurity [112].
Predicative analytics can further improve the integration and intelligence of security design. AI-
driven User Interface design must be user-centered instead of technology-centered [106]. The
complexity and the number of cyberattacks is increasing daily which poses risks to both public
andprivate technological assets [113] and must be ‘proactive’ and ‘balance’ risks with benefits
[114]. Documenting findings provides a record of the risks-benefits, which can be useful for
future reference or if questions arise later and to ensure that risks/benefits and strategies are not

16
forgotten or overlooked and sharing the findings with all relevant stakeholders, including
developers, managers and users ensures awareness and buy-in to achieve practical and effective
solutions [114]. This is the essence of the ‘transformative collaboration’ dimension. Examples of
AI-enableduser interface tools are Jarvis, Amazon Alexa, Netflix, IBM Watson, Nest Thermostat,
Spotify and iRobot Roomba [115]. User experience, designing of user interaction based on
human perception and minimizing the complexity of software integration will improve the
existing network infrastructure [116] and illustrate the principles of being ‘holistic’, ‘’integrative’
while enhancing the dimensions of ‘expertise building’ and creating a truly ‘AI-driven culture’. By
checking trainingdata sets and testing algorithms for bias, leaders can ‘proactively’ minimize AI
drawbacks [117]. “Organizations should not wait to implement AI technologies because they can
be strong tools in the fight against cybercrime” [118]. Drawing from various approaches from
different fields—Artificial Intelligence, Software Engineering and Computer-Human
Interaction— is ‘integrative’ and can increase the overall usability of devices [119].
Networks are persistently exposed to threats like malware, phishing, password breaches and
denial of service attacks. F-Secure Labs receives sample data of 500,000 files from its customers
that include 10,000 malware variants and 60,000 malicious URLs for analysis and protection daily.
Theyrely on AI and machine learning tools to detect suspiciousness based on the structure of a
file or its behavior or both [120]. In their survey, the top-performing risk mature companies
implemented twice as many of the key risk capabilities and generated three times the level of
performance as those in the bottom 20%. Companies that succeed in turning risks into results create
competitive ‘strategic’ advantage through efficient deployment of scarce resources, better
decision-making, reduced exposure to negative events and applying a broad ‘risk-benefit lens’ to
the business [50].
7. CONCLUSIONS
This research has systematically reviewed the most salient benefits and risks of AI and generative
AI in the current literature and categorized them by their impact on society. From this analysis,
examples of these challenges and benefits are categorized into 8 arenas based on recent relevant
research. Next, related risk models and theories, based upon risk models in IT project
management, cloud computing implementation and enterprise resource management and social
cognitive theory applied from psychology are synthesized into the context of AI resulting in the
newly proposed framework called “The Transformation Risk-Benefit Model of Artificial
Intelligence”.
This Model presents a comprehensive framework tailored to manage the ever-increasing changes
in the AI environment by integrating negative risks and positive benefits. The analysis of past
researchfound scant models and a hyper focus on risks to date. The context of AI was not fully
integrated into the previous empirically-based models. “The Transformation Risk-Benefit Model
of AI” expands the constructs of risk dimensions (such as ‘project size’ transforming into
‘resources’, ‘culture’ changing to ‘AI-driven culture’ and ‘external engagement’ widening to
‘transformational collaboration’ both internally and externally to the organization) based upon a
systematic review.
In addition to broadening the construct dimensions, this paper contributes new insight into both
the principles and processes which will enhance the transformation of risks into competitive
benefits and capacities.
To address the limitations and drawbacks of current risk models, the authors recommend working
on benefits and risks simultaneously using five processes of ‘assessing’ AI risks and
opportunities, ‘identifying’ the risks and opportunities, ‘analyzing’ diverse options, plans, and

17
strategies tied to these risks and benefits, ‘creating optimal plans’ through internal and external
collaboration and ‘evaluating and adjusting plans’ in a continuous improvement fashion. Next,
the authors add to the body of knowledge by presenting samples of pragmatic and innovative
solutions in each of the AI and generative AI risk dimensions that highlight how risk can
transform into benefit in the latest AItechnology context. Lastly, three use cases in healthcare,
climate change/environment and cyber security add depth to understanding the AI context and
illustrate how each sector can uniquely address the principles, dimensions and processes to
transform AI risks into benefits overall.
REFERENCES
[1] Cheatham, B.,Javanmardian, K. &Hamid Samandari, H. (2019, April 26)“Confronting the risks
ofartificial intelligence.”McKinsey Quarterly.
[2] Dwivedi, Y. K., Hughes, L., Ismagilova, E., Aarts, G., Coombs, C., Crick, T., Duan, Y., Dwivedi,
R., Janssen, M. et al. (2021). “Artificial Intelligence (AI).” International Journal of Information
Management. Vol. 57, doi:10.1016/j.ijinfomgt.2019.08.002
[3] Iliescu, A.B. & Bustos, A. (2023, March 29)“How using AI can benefit society?”AI for Good.
[4] Vinuesa, R., Azizpour, H., Leite, I., Balaam, M., Dignum, V., Domisch, S., Felländer, A.,
Langhans, S. D., Tegmark, M. &Nerini, F. F. (2020) “The role of artificial intelligence in achieving
the Sustainable Development Goals.” National Communications, Vol. 11, pp1-233.
[5] Manyika, J., Lund, S. Chui, M., Bughin, J., Woetzel, J., Batra, P., Ko, R. & Sanghvi, S. (2017,
November 28) “Jobs lost, jobs gained: What the future of work will mean for jobs, skills, and
wages?’McKinsey Global Institute.
[6] Jonsson, A. & Svensson, V. (2016) “Systematic lead time analysis.”Chalmers University of
Technology Publications.
[7] Curry, Rachel (2023, August 10) “How AI can help create jobs for humans, not just automate
them.” Technology Executive Council.
[8] Monahan, Kelly (2023, June 27) “Generative AI disconnect.” Upwork.
[9] Rubik, B. &Jabs, H. (2018) “Artificial Intelligence and the biofield: New opportunities and
challenges.” Cosmos and History, Vol. 14, No. 1, pp153-162.
[10] Bergquist, T., Yan, Y., Schaffter, T., Yu, T., Pejaver, V., Hammarlund, N., Prosser, J., Guinney, J.
& Mooney, S. (2020) “Piloting a model-to-data approach to enable predictive analytics in health
care through patient mortality prediction.” Journal of the American Medical Informatics
Association: JAMIA, Vol. 27, No. 9, pp1393-1400.
[11] Harada, T., Miyagami, T., Kunitomo, K. & Shimizu, T. (2021, August 10) “Clinical Decision
Support Systems for diagnosis in primary care: A scoping review.” International Journal of
Environmental Research and Public Health, Vol. 18, doi:10.3390/ijerph18168435
[12] Metzger, M. & Donahoe, E. (2019) “Artificial Intelligence and human rights.” Journal of
Democracy, Vol. 30, No. 2, pp115-126.
[13] Zheng, Y. & Hu, X. (2020) “Healthcare predictive analytics for disease progression: A longitudinal
data fusion approach.” Journal of Intelligent Information Systems, Vol. 55, pp351–369.
[14] OECD correspondent, (2019) “Artificial Intelligence in society.”OECD Publishing.
[15] Data Dome Correspondent, (2021, September 18) “How AI is used in fraud detection,” Data Dome,
[16] Gregory, J. G. (2023, Sept 23) “How data scientists fight deep fakes in cyberspace.” Nutanix,
[17] Zulaikha, S., Mohamed, H., Kurniawati, M. Rusgianto, S. & Rusmita. S. (2021) “Customer
predictive analytics using Artificial Intelligence.” The Singapore Economic Review, Vol. 66, No. 2,
pp1-12.
[18] Stoica, A.A. & Chaintreau, A. (2019) “Hegemony in social media and the effect of
recommendations.” The Web Conference 2019, Vol. 2, pp575-580.
[19] Abarca-Alvarez, F., Campos-Sanchez, F., Reinoso-Bellido, R. (2018) “Demographic and dwelling
models by Artificial Intelligence: Urban renewal opportunities in Spanish Coast.” International
Journal of Sustainable Development and Planning, Vol. 13, No. 7, pp941-953.
[20] Shukla, N, Tiwari, M. & Beydoun, G. (2019) “Next generation smart manufacturing and service
systems using big data analytics,” Computers & Industrial Engineering, Vol. 128, pp905-910.
[21] Spanaki, K., Gürgüç, Z, Adams, R & Mulligan, C (2018), “Data supply chain (DSC),” International
Journal of Production Research, Vol. 56, No.13, pp4447-4466.

18
[22] Olshannikova, E., Ometov, A., Koucheryavy, Y. & Olsson, T. (2015) “Visualizing Big Data with
augmented and virtual reality,” Journal of Big Data, Vol. 2, pp22-49.
[23] Zhong, Xu, Chen & Huang.(2015)“Big Data Analytics for physical internet-based intelligent
manufacturing shop floors,” Inter’ll Journal of Production Research, pp1-14.
[24] Ciucci, M. & Gouardères, F. (2020, April) “Briefing, requested by the ITRE Committee, European
Parliament, white paper on Artificial Intelligence”, PE 648.773, Policy Department for Economic,
Scientific and Quality of Life Policies, doi: 10.2861/614816
[25] Mergel, I. (2018) “Open innovation in the public sector: Drivers and barriers for the adoption of
challenges,” Public Managerial Review, Vol. 20, pp726–745.
[26] Mikhaylov, S. J., Esteve, M. & Campion, A. (2018). “Artificial intelligence for the public sector.”
Philosophical transactions. Series A, Mathematical, physical, and engineering sciences Vol. 376,
No. 2128, e20170357. doi:10.1098/rsta.2017.0357.
[27] Eggers, W. & Bellman, J. (2015) “The journey to government's digital transformation,”Deloitte.
[28] Korinek, A. & Stiglitz, J. E. (2020, January 23) "Artificial Intelligence and its implications for
income
[29] distribution and unemployment." Chapter 14, pp349-390 In The Economics of Artificial
Intelligence: AnAgenda. Agrawal, Gans and Goldfarb (Editors). University of Chicago Press,
[30] Andreassen, CS. (2015) “Online social network site addiction: A comprehensive review.” Current
Addiction Reports, Vol. 2, No. 2, pp175–184.
[31] Ajunwa, I. (2021) “An auditing imperative for automated hiring.” Harvard Journal of Law &
Technology, Vol. 34, No. 2, pp1-66.
[32] Barocas, S., Hardt, M. & Narayanan, A. (2018) “Fairness and Machine Learning: Limitations and
opportunities. Available: http://fairmlbook.org
[33] Agarwal, S., Farid, H., Gu, Y. , He, M., Nagano, K. & Li, H. (2019) “Protecting world leaders
against deep fakes.” In Proceedings of the IEEE Conference on Computer Vision and Pattern
Recognition (CVPR) Workshops. IEEE, Long Beach, CA, pp38-45.
[34] Kreps, S., McCain, R. M. & Brundage, M. (2020) “All the news that’s fit to fabricate: AI-generated
text as a tool of media misinformation. Journal of Experimental Political Science, pp1–14.
[35] Ledwich, M. & Zaitsev, A. (2020) “Algorithmic extremism: Examining YouTube’s rabbit hole of
radicalization.” First Monday, Vol. 25, No. 3, doi:10.5210/fm.v25i3.10419.
[36] Larson, J. Mattu, S., Kirchner, L. & Angwiny, J. (2016, May 23) “How we analyzed the
COMPASrecidivism algorithm,”ProPublica. Available
[37] Selbst, A.D. (2017, February 25) “Disparate impact in big data policing,” Georgia Law Review,
Vol. 52, No. 1, pp109–195.
[38] Dastin, J. (2018, October 10) “Amazon scraps secret AI recruiting tool that showed bias against
women,” Rueters.
[39] Hao, K. (2020, August 20) “The UK exam debacle reminds us that algorithmscan’t fix broken
systems,”MIT Technology Review.
[40] Cook, J. (2021, Nov. 3) “Why the ibuying algorithms failed Zillow, and what it says about the
business world’slove affair with AI.”GeekWire.
[41] Rapier, G. (2019, April 1) “Hackers steered a Tesla into oncoming traffic by placing 3 small stickers
on the road.”Business Insider.
[42] Baxter, K. & Schlesinger, Y. (2023, June 6) “Managing the risks of generative AI.”Harvard
Business Review.
[43] Aziz, S. & Dowling, M. (2019) “Machine learning and AI for Risk Management,” in Chapter 3
pp33-50, Disrupting Finance: FinTech and Strategy in the 21st Century. Lynn, J Mooney, Rosati
and Cummins, (Eds.) Palgrave Studies in Digital Business and Enabling Technologies. Palgrave
Pivot.
[44] Wilson, H. J., Daugherty, P. & Bianzino, N. (2017) “The jobs that artificial intelligence will create.”
MIT Sloan Management Review, Vol. 58, No. 4, pp14–16.
[45] Illanes, P., Lund, S., Mourshed, M., Rutherford, S. & Tyreman, M. (2018, January 22) “Retraining
and reskillingworkers in the age of automation,”McKinsey Global Institute.
[46] McFarlan, F. W. (1981) “Portfolio approach to information systems.” Harvard Business Review,
Sept – Oct, pp142-150.
[47] Dey, P. K., Clegg, B. & Cheffi. W. (2011) “Risk management in enterprise resource planning
implementation:
[48] A new risk assessment framework,” Production Planning & Control,

19
[49] Sarens, G. & Christopher, J. (2010) “The association between corporate governance guidelines and
risk management and internal control practices.”Managerial Auditing Journal, Vol. 25, No. 4,
pp288-308.
[50] Ganesh, A. D. & Kalpana, P. (2022) “Future of Artificial Intelligence and its influence on supply
chain risk management.” Computers & Industrial Engineering, Vol. 169, e108206.
[51] Locascio, L.A. (2023, January 26) “NIST Risk Management Framework Aims to Improve
Trustworthiness ofArtificial Intelligence,” pp1-48, National Institute of Standards and Technology
(NIST), U.S. Department of Commerce.
[52] van Kessel, P., Layman, J., Blackmore, J., Burnet, I. & Haroda, S. (2013) “Turning risk into
results,” Ernst and Young Global Limited, pp1-16.
[53] Yip, A, Brehm, L. & Wagner, C. (2017) “Revisiting McFarlan’s Risk Model for IS implementation
success -Does culture matter?”Proceedings of the 28th Australasian Conference on Information
Systems, Hobart, Australia, Vol. 32, May 12.
[54] Elkington, J. (1994). "Towards the sustainable corporation: Win-Win-Win business strategies for
sustainable development," California Management Review, Vol. 36, No. 2, pp90–100.
[55] Bandura, A (1989) “Human agency in social cognitive theory.” American psychologist, Vol. 44,
No. 9, pp11751186.
[56] Ratten, V. (2016) “Continuance use intention of cloud computing: Innovativeness and creativity
perspectives,” Journal of Business Research, Vol. 69, No. 5, pp1737-1740.
[57] Srinivas, K. (2019) “Process of risk management” April, Chapter 11 in Hessami, A.G. (Ed.),
Perspectives on Risk, Assessment and Management Paradigms, InTechOpen, .
[58] Aloini, D., Dulmin, R., Mininno, V. & Diotisalvi, V. (2007) “Risk management in ERP project
introduction: Review of the literature.” Information & Management, Vol. 44, July 12, pp547–567.
[59] [57} Tzanakakis, K. (2021) “Risk assessment techniques.” In: Managing Risks in the Railway
System. Springer Tracts on Transportation and Traffic, Chapter 4, Vol 18, pp113-146.
[60] Bridges, J., (2023). “Project Risk Analysis: Tools, templates & techniques.”Project Manager.
[61] Torabi, S.A., Giahi, R. & Sahebjamnia, N. (2016) “An enhanced risk assessment framework for
businesscontinuity management systems.”Safety Science, Vol. 89, pp201-218.
[62] ISO-31010:2009, (2009) “Risk management - Risk assessment techniques.” International
Organization for Standardization, Available:https://www.iso.org/standard/51073.html
[63] UCLA Research Administration. (2020) “Conducting risk-benefit assessments and determining
level of IRB review.” November 30, Available: https://ohrpp.research.ucla.edu/assessing-risks/
[64] ISO-22301:2012. (2012) “Business continuity management”International Organization for
Standardization.
[65] ISO-31000:2018, (2018) “Risk management: A practical guide.”International Organization for
Standardization.
[66] International Atomic Energy Agency (IAEA), (2001) “Risk management: A tool for improving
nuclear powerplant performance,” IAEA-TECDOC-1209, IAEA, Vienna.
[67] Heckmann, I., Comes, T., & Nickel, S. (2015) “A critical review on supply chain risk—Definition,
measure and modeling.” Omega, Vol. 52, pp119–132.
[68] Lockamy, A. (2011), "Benchmarking supplier risks using Bayesian networks", Benchmarking: An
International Journal, Vol. 18 No. 3, pp409-427.
[69] Schwab, L., Gold, S. & Reiner, G. (2019) “Exploring financial sustainability of SMEs during
periods of production growth,” International Journal of Production Economics, Vol. 212, pp8–18.
[70] Gerasimova, E. & Redin, D. (2015) “Analyzing and managing financial sustainability of the
company in turbulent environment.” Mediterranean Journal of Social Sciences, Vol. 6, No. 4,
pp138-143.
[71] EMA Team (2011) “Benefit-risk methodology project Work package 1 report: Description of the
current practiceof benefit-risk assessment for centralized procedure products in the EU regulatory
network” May 25, Europeans Medicine Agency: EMA/227124/2011.
[72] Hefter, A., Loadwick, O., T Poppensieker, T. & Raufuss, A. (2022) “Risk transformation: The heart,
the art, andthe science.”McKinsey & Company, April, pp1-10.
[73] K2 Partners. (2022, July 7) “Limitations in traditional risk management and why you should switch
toenterprise.”Insights.
[74] Claypole, A. (2021, June 8) “Traditional risk management vs enterprise risk management: Which
approach isbest?”Ideagen.

20
[75] Witcher, M. F. (2021). “Using system risk structures to understand and balance risk/benefit trade-
offs.” April 23, Outsourced Pharma. [74] Cooper, A.F., Moss, E., Laufer, B. & Nissenbaum, H.
(2022) “Accountability in an algorithmic society:
[76] Relationality, responsibility, and robustness in machine learning,” in 2022 ACM Conference on
Fairness, Accountability, and Transparency Proceedings, June 21-24, 2022, Seoul, Republic of
Korea, ACM, New York, NY, USA, pp864–876.
[77] Anderson, J., Rainie, L. & Luchsinger, A. (2018, December 10) “Artificial Intelligence and the
future of humans.” Pew Research Center, pp1-116.
[78] Sabhlok, S. (2022, January) “Introducing AI factsheets on cloud packfor data as a service: Automate
collectionof model facts across the AI Lifecycle.” Community IBM.
[79] Kulynych, B., Overdorf, R., Troncoso, C. & Gürses, S. (2020) “POTs: Protective Optimization
Technologies.” Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency.
January 27-30, 2020, Barcelona, Spain, ACM, New York, NY.
[80] Miclaus, K. (2022, July 29) “Top 3 barriers to entry for AI in Health Care.” Dataiku.
[81] Wang, Y., Kung, L., Wang, Y. C.W. & Cegielski, C.G. (2018) “Integrated big data analytics-
enabled transformation model,” Information and Management, Vol. 55, No. 1, pp64-79.
[82] Jamshidi, M. Lalbakhsh, A. Talla, J. , Peroutka, Z. , Hadjilooei, F. , Lalbakhsh, P. , Jamshidi, M.,
La Spada, L., Mirmozafari, M., Dehghani, M. , Sabet, A., Roshani, S., Roshani, S., Bayat-Makou,
N., et. al. (2020) “AI and COVID-19: Deep learning approaches for diagnosis and treatment.” IEEE
Access, Vol. 8, pp1-15.
[83] Abdulkareem1, M. & Petersen, S.E. (2021) “The promise of AI in detection, diagnosis, and
epidemiology forcombating COVID-19” Frontiers in Artificial Intelligence, Vol. 4, May 21.
[84] Gupta, N.S. & Kumar, P. (2023) “Perspective of Artificial Intelligence in healthcare data
management,” Computers in Biology and Medicine, Vol. 162, 107051.
[85] Adam, B. & Memarzadeh, K. (2020) “The rise of Artificial Intelligence in healthcare applications.”
Artificial Intelligence in Healthcare, pp25–60.
[86] Cirillo, D. & Alfonso Valencia, A. (2019) “Big data analytics for personalized medicine,” Current
Opinion in Biotechnology, Vol. 58, pp161-167.
[87] Wang, Y., Kung, L.& Byrd, T.A. (2018) “Big data analytics,” Technological Forecasting & Social
Change, Vol. 126, No. C, pp3-13.
[88] Kornelia, B. & Ślęzak, A. (2022) “The use of Big Data Analytics in healthcare.” Journal of Big
Data, Vol. 9, No.1, pp3-27. doi:10.1186/s40537-021-00553-4
[89] Kasaw, A.,Gizachew, M. &Kendie, S. (2019) “The role of medical data in efficient patient care
delivery: A review.” Risk Management and Healthcare Policy, Vol. 12, April 24, pp67-73.
[90] Wallace, David K. (2010) “Evidence-based medicine and levels of evidence.” The American
Orthopedic Journal, Vol. 60. pp2-5. doi:10.3368/aoj.60.1.2
[91] Wallace, P., Shah, N.,Dennen, T. Bleicher, P. &Crown, W.(2014) “Optum Labs: Building a novel
node in the learning health care system.” Health affairs, Vol. 33, No. 7, pp1187-94.
[92] Raghupathi, W. & Raghupathi, V. (2014) “Big data analytics in healthcare: Promise and potential.”
Health Information Science and Systems, Vol. 2, No. 3. Feb. 7.
[93] Simpao, A.F., Ahumada, L.M. & Rehman, M.A. (2015) “Big data and visual analytics in
anaesthesia and health care,” BJA: British Journal of Anesthesia, Vol. 115, No. 3, pp350–356
[94] McLaughlin, K.A., Weissman, D, & Bitran, D. (2019) ”Childhood adversity and neural
development: A systematic review.” Annual Review of Developmental Psychology, Vol.1, pp277-
312.
[95] Cox, M., Cuff, P., Brandt, B., Reeves, S. & Zierler, B. (2016) “Measuring the impact of
interprofessional education on collaborative practice and patient outcomes.” Journal of
Interprofessional Care, Vol. 30, No. 1, pp1-3.
[96] Alqahtani, T., Badreldin H., Alrashed, M., Alshaya, A., Alghamdi, S., Saleh, K., Shuroug, A., et. al
(2023).
[97] “The emergent role of Artificial Intelligence, natural learning processing, and large language
models in higher education and research.” Research in Social and Administrative Pharmacy, Vol.
19, No. 8, pp1236-1242.
[98] Harrison, B., Stall, S. , Cutcher-Gershenfeld, J., Vrouwenvelder, K., Wirz, C., Rao, Y. &Peng, G.
(2023). “Garbage in, garbage out: Mitigating risks and maximizing benefits of AI in research.
Nature, Vol. 623, November 2, pp28-31.

21
[99] Nature Editorial Group, (2022) “Pervasive machine learning in physics.” Nature Reviews Physics,
Vol. 4, pp353-354. doi:10.1038/s42254-022-00475-x
[100] Hilburn, K.A., Ebert-Uphoff, I. & Miller, S.D. (2021) “Development and interpretation of a neural-
networkbased synthetic radar reflectivity estimator using GOES-R satellite observations.” Journal of
Applied Meteorology and Climatology, Vol. 60, No. 1, pp3-21.
[101] Schneider, T., Behera, S., Boccaletti, G., Deser, C., Emanuel, K., Ferrari, R., Leung, L., Lin, N.,
Müller, T., Navarra,A., Ndiaye, O., Stuart, A., et.al (2023) “Harnessing AI and computing to
advance climate modelling and prediction.”Nature Climate Change, Vol. 13, pp887–889.
[102] Rolnick, D. et al. (2022) “Tackling climate change with machine learning.” ACM Computing
Surveys, Vol. 55, Issue 2, No 42, pp1–96.
[103] Ryan-Mosley, T. (2023). “How AI can actually be helpful in disaster response.” The Technocrat,
MIT Technology Review, February 20.
[104] Oprihory, J.L. (2020) “California National Guard Using Satellites to Fight Wildfires.” Air and
Space Forces Magazine, September 8.
[105] Pandey, S. (2017) “Flooding affects millions in Bangladesh, India and Nepal.” UNICEF, August
21.
[106] Inanc, Y. S. (2023) “Turkey earthquake: Time stands still in Adiyaman but anger grows overslow
response.” Middle East Eye, February 16.
[107] Kaack, L., Donti, P., Strubell, E., Kamiya, G., Creutzig, F. & Rolnick. (2022) “Aligning AI with
climate change mitigation.” Natural Climate Change, Vol. 12, pp518–527.
[108] Dash, B., Ansari, M. F., Sharma, P. & Ali, A. (2022) “Threats and opportunities with AI-based
cyber security intrusion detection: A review.” International Journal of Software Engineering &
Applications (IJSEA), Vol.13, No. 5, pp13-21.
[109] Akinsola, J. E. T., Adeagbo, M., Kayode, O., Akinsehinde, S. & Onipede, F. (2022) ”Artificial
Intelligence emergence in disruptive technology.” In Salau, A.O., Jain, S. & Sood, M. (Eds),
Computational Intelligence and Data Sciences, CRC Press, pp63-90.
[110] Nwiabu, N., Allison, I., Holt, P., Lowit, P. & Oyeneyin, B. (2012) “User interface design for
situation-aware decision support systems.” In: IEEE Multi-disciplinary Conference on Cognitive
Methods in Situation Awareness and Decision Support. 6-8 March. Piscataway, New Jersey: IEEE.
pp332-339.
[111] Samsonov, P. A. (2016) “Improving interactions with spatial context-aware services.” In
Companion Publication of the 21st International Conference on Intelligent User Interfaces (IUI '16
Companion). Association for Computing Machinery, New York, NY, USA, pp114-117.
[112] Zhai. S. (2017) “Modern touchscreen keyboards as intelligent user interfaces: A Research review.”
In Proceedings of the 22nd International Conference on Intelligent User Interfaces (IUI '17).
Association for Computing Machinery, New York, NY, USA, pp1-2. doi:
10.1145/3025171.3026367
[113] Akinsola, J., Kuyoro, A., Adeagbo, M. & Awoseyi, A.A. (2020) “Performance evaluation of
software using formal methods.” Global Journal of Computer Science and Technology,” Vol. 20, ,
pp17–23.
[114] Quinn. P. & Zhai, S. (2016) “A cost-benefit study of text entry suggestion interaction.” Proceedings
of the 2016 Conference on Human Factors in Computing Systems (CHI '16), May 2016, pp83–88.
[115] Mansourian, A. (2023) “Data privacy and cybersecurity regulations introduction.” NMS Consulting.
[116] EC-Council, (2023) “What is cybersecurity?”
[117] Gonzalez, C. (2023) “Threat modeling: 5 steps, 7 techniques and tips for success.”Info Sec Trends.
[118] Mandava, S.K. (2020) “User interfaces with Artificial Intelligence,” Artificial Intelligence, User
Interface, User Experience, Vol. 38.
[119] Strehlow, R. (2018, March 11). “Cyber security requires an important ingredient,”Strong UX.
[120] Parish, K. (2023, April 18) “AI and UX Design: How AI can enhance user experience.”
[121] Wlosinski, L.G. (2020, January 1) “Understanding and managing the Artificial Intelligence threat,”
ISACA Journal, Vol. 1, pp 40-43.
[122] Jalil, N. (2022) “Introduction to Intelligent User Interfaces (IUIs)”, in Castro, N. Cabrero, D. &
Heimgartner, R. (Eds.), Software Usability. IntechOpen, doi: 10.5772/intechopen.97789.
[123] Press Trust of India, (2017) “Cyber security firms turn to Artificial Intelligence as hacking threats
rise.” ND TV Profit, November

22
AUTHORS
Professor Richard A. Fulton (M.S., Illinois State University)has taught full time
computer science and information systems courses at Troy University – e campus for the
past 15 years and previously at Illinois State University. His articles have been published
in The Journal of Technology Research, The Journal of Scientific Information on Political
Theory, Developments in Business Simulations and Experiential Learning, and the
International Journal of Innovation, Technology and Management.
Dr. Diane J. Fulton (Ph.D., University of Tennessee-Knoxville) is Emeritus Professor of
Management at Clayton State University, located in Morrow, Georgia. Her research
interests include advanced technologies, innovations and online teaching tools. She has
published several books, book chapters, and numerous articles in academic journals,
including California Management Review, Planning Review, Journal of Small Business
Management, International Journal of Management Education and Entrepreneurship
Theory
Susan Kaplan (BAS, MAS, University of Minnesota-Dultuh) is Executive Vice President
and Chief Management Officer of Modal Technology Corporation, a high-tech firm
located in Minneapolis, Minnesota that offers new and proven solutions for artificial
intelligence and machine learning that are based on modal interval arithmetic. Ms. Kaplan
was Founder and President for Quality Management Solutions. She led and restructured
organizations in healthcare, government, manufacturing, and service sectors to improve
profitability. She is the author of The Grant Writing Process.
Nathan Hayes, CEO and Founder at Modal Technology Corporations, is an entrepreneur,
mathematician, and software architect with more than 20 years of experience working in
these combined fields. Hayes specializes in the applied science of modal interval analysis
to the fields of artificial intelligence, machine learning and high performing computing.

THE TRANSFORMATION RISK-BENEFIT MODEL OF ARTIFICIAL INTELLIGENCE:BALANCING RISKS AND BENEFITS THROUGH PRACTICAL SOLUTIONS AND USE CASES

More Related Content

Similar to THE TRANSFORMATION RISK-BENEFIT MODEL OF ARTIFICIAL INTELLIGENCE:BALANCING RISKS AND BENEFITS THROUGH PRACTICAL SOLUTIONS AND USE CASES (20)

Recently uploaded (20)

THE TRANSFORMATION RISK-BENEFIT MODEL OF ARTIFICIAL INTELLIGENCE:BALANCING RISKS AND BENEFITS THROUGH PRACTICAL SOLUTIONS AND USE CASES