Bold, Brave, & Benevolent: A Call to Action for Leaders On Active Inference, Gray Zone Conflict, and AI Issues We're Overlooking

Senior leaders in both the private and public sector face increasing challenges in understanding and navigating AI methods, leadership, and governance due to the rapid pace of technological change, the blurring lines between authentic and AI-generated interactions, and the need to adapt traditional governance structures to address complex, interconnected threats in an era of widespread internet access and advanced generative AI capabilities. Today, I'd like to share some reflections on three topics that I believe are shaping our future. My hope is that these ideas resonate with readers and spark some interesting conversations.

1. Active Inference: A Promising Path Beyond GenAI

The first topic is one that builds on themes that I've been emphasizing for about 6 years, including an MIT Sloan Management Review article friend R "Ray" Wang and I co-authored in 2019. Specifically: deep learning isn't the only way to do AI and it has both positive and negatives. In early 2024 I had the opportunity to serve as host and facilitator for five dinners at 2024's World Economic Forum meetings in Davos on the topic of active inference methods applied to healthcare, smart cities, space, supply chains, and policy activities. At the time it was clear that the benefits of Active Inference relative to Deep Learning were not fully understood either by leaders in the private or public sectors. Now however it seems like that's starting to be heard - if still as an early signal on the horizon. Specifically, Active Inference and inference-based methods for AI offer several key advantages over deep learning:

• Adaptability to Novelty: Unlike current GenAI systems that struggle with situations not represented in their training data, Active Inference can adapt more readily to new scenarios. This is crucial in today's rapidly changing business and geopolitical environment, where unprecedented challenges are becoming the norm.
• Efficiency: Active Inference is generally more efficient in terms of computational resources and energy consumption. This aligns well with corporate sustainability goals and can lead to significant cost savings, especially as AI becomes more pervasive in business operations.
• Transparency: Active Inference systems have the potential for more transparent decision-making processes. In an era of increased scrutiny on AI ethics and decision-making, this can help build trust with stakeholders and potentially mitigate legal and reputational risks.
• Localized and Predictive Capabilities: As I am regularly reminding folks in keynotes and other talks I give, Active Inference may prove better for more predictive, localized AI, especially in situations where the present is not like past data. This is particularly valuable in rapidly evolving markets or during crisis situations.
• Reduced Data Dependency: While current GenAI using Deep Learning is massively data and CPU intensive, Active Inference potentially can operate with less historical data. This could be a game-changer for organizations dealing with limited data sets or in rapidly changing environments where historical data quickly becomes obsolete.
• Reduced GPU Demands and Associated Energy Dependency: Unlike current GenAI systems that are massively data and CPU intensive, Active Inference can potentially operate more efficiently in terms of computational resources and energy consumption, aligning well with corporate sustainability goals and potentially leading to significant cost savings.
• Biological Plausibility: Active Inference aligns more closely with biological cognitive processes. This could lead to more intuitive AI systems that better mimic human decision-making processes, potentially leading to more natural human-AI collaboration.

It is absurd for the world to massively increase its energy demands for GenAI when active inference offers a more sustainable and efficient alternative. Current deep learning-based GenAI systems are environmentally unsustainable and economically shortsighted due to their intensive data and CPU requirements. Active inference presents a superior approach that is less data-intensive, less GPU-intensive, and less energy-intensive. Instead of investing in power-hungry infrastructure, we should focus on these advanced AI methods.

In addition, the promise of Active Inference to handle novelty and uncertainty will be crucial in today's rapidly changing business environment. Unlike current GenAI systems, which struggle with situations not represented in their training data, Active Inference as an AI method will adapt more readily to new scenarios. This adaptability is invaluable for executives dealing with unprecedented market conditions, emerging competitors, or disruptive technologies. My sense is, over the next 12-18 months, by adopting inference-based AI methods, we can align technological progress with sustainability goals, reduce costs, and mitigate risks associated with AI ethics and decision-making transparency. This shift would benefit organizations and contribute to global efforts to combat climate change (by reducing energy demand for AI) while simultaneously promoting more responsible AI innovation.

Lastly on this specific topic regarding additional methods to do AI - I'd like to extend a special recognition to France Hoang, co-founder of BoodleBox.ai, for his pioneering work in making AI accessible to educational environments and organizational settings seeking to work across different AI bots. France's efforts at BoodleBox exemplify the kind of innovative, cross-sector collaboration we need to harness AI's potential for positive societal impact.

2. Reflections on Leadership and Governance

As a second topic, I recently had the pleasure of participating in an engaging podcast conversation with Meghan Day and Dottie Schindlinger of the Diligent Institute & Co-Host of the Corporate Director Podcast. I'd like to share some of that discussion here specifically on critical issues that senior executives need to grapple with in today's complex business environment.

The growing challenge of distinguishing between authentic human interactions and AI-generated content is not just a technical issue, but a strategic one. Executives need to be aware of how this impacts everything from customer interactions to competitive intelligence. I shared some of these challenges, including one where I was part of a team that faced more 23 million bot-manufactured comments in 2017, to underscore the scale of this challenge and how it has only grown more complex over time. I also raised with Meghan the challenge of "gray zone conflict" (more on this topic in the closing part of this post), specifically situations that fall short of outright war but still pose significant threats, as being particularly relevant for executives in both the private and public sectors for our interconnected world. Cyber threats, media manipulation campaigns, and economic warfare can all fall into this category, and executives need to be prepared to navigate these murky waters.

The need to upgrade governance models is perhaps one of the most crucial takeaways for senior executives. Traditional governance structures are often ill-equipped to handle the pace of technological change and the complexity of modern threats. Executives need to foster environments that welcome diverse perspectives and encourage innovation, while still maintaining necessary controls. In chatting with Meghan, I suggested the power of simple yet important questions that leaders should ask to encourage team members to tackle problems together - including "How are things going?", "How are people feeling?", and "What motivates you and brings you joy?" These questions can be transformative tools for executives and help uncover uncover hidden issues, boost employee engagement, and drive innovation from within the organization.

Many thanks to both Dottie and Meghan for the opportunity to join this important discussion on the essential topics of modern governance and leadership in turbulent times. During the recording process, I smiled when they said they had a hard time striving to encapsulate the different range of experiences in my life - and really enjoyed the breadth and depth of our discussion with the full podcast.

3. Addressing the Challenges of GenAI in Free Societies

As a third and final topic, I'd like to share a summary of the slides I presented during an opening keynote at a University of Tennessee event around increasing gray zone conflict and the impacts of GenAI. While the specifics of the conference and discussion points were non-attribution, I'm comfortable highlighting the key points I raised in my presentation since it was in my voice and it echoes earlier presentations I have given where the slides can be found online, including one with NSI and the Department of Defense's Strategic Multilayer Assessment (SMA) program activity on the topic of "Why, What, and How Free Societies Must Develop Effective Deterrence of Actors Intending Bad Ends Given Our GenAI Era".

The keynote I gave at the University of Tennessee a few weeks ago covered a range of critical topics at the intersection of technology, policy, and societal challenges. I discussed the rapid advancement of GenAI and its implications relative to increasing conflicts within and across civil societies. The convergence of data proliferation and globalization poses new challenges for data governance and privacy. In our current GenAI era, Executives need to be acutely aware of the risks associated with data hoarding, both within their own organizations and by external entities. This trend can impact everything from competitive strategy to regulatory compliance.

The escalating frequency, sophistication, and cost of ransomware and cyber threats should be a top concern for executives across all industries - especially GenAI is advancing the sophistication of these scams and attacks. As the attack surface expands with the growth of network devices, the potential impact of a successful attack grows exponentially. This underscores the need for robust cybersecurity strategies and continuous risk assessment.

Professionally I believe executives in both the private and public sector need to focus understanding the motives and behaviors of the human operators behind AI, rather than just the technology itself. The potential for GenAI to be exploited by authoritarian regimes or actors with values divergent from more free and representatives forms of civil societies represents a significant geopolitical risk that executives need to be aware of, particularly for companies operating globally. This could impact everything from supply chain security to market access in certain regions.

In conclusion, the challenges presented by GenAI and increasing gray zone conflicts require a multifaceted approach that goes beyond traditional cybersecurity measures. We must work together to develop new forms of deterrence that focus on the human actors behind GenAI abuses, rather than just the technology itself. This includes democratizing the tradecraft of "information discernment" and promoting "digital dignity" across society, while maintaining consistency with constitutional principles.

Going Forward and a Call to Action

Together we must lead the charge in creating robust, whole-of-society solutions that improve our ability to identify and mitigate GenAI abuses, update privacy protections, and establish clear liability regimes. Moreover, working across sectors we need to empower all of us - everyone in civil society - with tools for information discernment and create trusted, non-political spaces to operationalize these efforts. As leaders in the private and public sectors, we must embrace better approaches to AI that do not erode privacy, do not exponentially grow our energy consumption, and provide more explainable approaches by design.

Hat-tip to Eric Prostejovsky & Cliff Sundstrom of GovCon Different for capturing this Call to Action so succinctly and skillfully on a brief 1 min 45 second video which if you haven't seen it is here. Ultimately to be effective positive #ChangeAgents in this era of rapid technological advancement, we must strive to create robust, whole-of-society solutions that harness AI's potential for the greater good, preserve our values, and promote digital dignity – let us be bold, brave, and benevolent in facing these challenges together!