![AI is here!
Preparing for a sea change
Surf with solid fill
September 24, 2024
[M]](https://image.slidesharecdn.com/04aiastoolsinyourcareer-241204190712-2e35aced/85/Exploring-AI-as-tools-in-your-career-pdf-2-320.jpg)
![Full Agenda
• Introductions
• AI in Perspective
• AI Concepts
• The Field of AI
• AI in Your Life and Career
• AI in Your Job Search
• AI in Your Business
• Show & Tell
• Q&A
Indicates a demo
is available
[M] Mentorship Session
[M]
[M]
[M]
[M]](https://image.slidesharecdn.com/04aiastoolsinyourcareer-241204190712-2e35aced/85/Exploring-AI-as-tools-in-your-career-pdf-3-320.jpg)
![Company
Websites &
Applications
Custom
ChatGPT
Data Collection
& Preparation
Indexing &
Vectorization
[ AI Search +
Semantic
Ranking ]
OpenAI API
Chat &
Completion
ChatBot
WebApp
Custom
Code
Manual
Data
Export or
Scheduled
API Pull
ChatBot UI
Users
OpenAI
Customer Techmien/Azure
Corporate data
Complementary data
GPT-friendly data
Search-friendly data
Index & vector data
](https://image.slidesharecdn.com/04aiastoolsinyourcareer-241204190712-2e35aced/85/Exploring-AI-as-tools-in-your-career-pdf-43-320.jpg)
![AI in Your
Life & Career
AI will not displace workers;
workers that use AI
will displace those that don’t
[M]](https://image.slidesharecdn.com/04aiastoolsinyourcareer-241204190712-2e35aced/85/Exploring-AI-as-tools-in-your-career-pdf-49-320.jpg)
![Using AI
in Your Job Search
[M]](https://image.slidesharecdn.com/04aiastoolsinyourcareer-241204190712-2e35aced/85/Exploring-AI-as-tools-in-your-career-pdf-57-320.jpg)
![Q & A
[M]](https://image.slidesharecdn.com/04aiastoolsinyourcareer-241204190712-2e35aced/85/Exploring-AI-as-tools-in-your-career-pdf-94-320.jpg)
Exploring AI as tools in your career.pdf
- 1. AI is Here!
- 2. AI is here! Preparing for a sea change Surf with solid fill September 24, 2024 [M]
- 3. Full Agenda • Introductions • AI in Perspective • AI Concepts • The Field of AI • AI in Your Life and Career • AI in Your Job Search • AI in Your Business • Show & Tell • Q&A Indicates a demo is available [M] Mentorship Session [M] [M] [M] [M]
- 4. Land Acknowledgement We acknowledge that the land we are meeting in is the traditional territory of many nations, including the Mississaugas of the Credit, the Anishinaabeg, the Chippewa, the Haudenosaunee and the Wendat peoples, and is now home to many diverse First Nations, Inuit and Métis peoples. We also acknowledge that Toronto is covered by Treaty 13 with the Mississaugas of the Credit. Land Acknowledgement
- 5. Introductions • Carlos E. Paz-Soldan • BSc. Industrial & Systems Eng. UNI (Lima, Peru) • MBA UofT Rotman’s • https://www.linkedin.com/in/carlospazsoldan • Founder & past CEO, Tenet Computer Group Inc. • 1984 - 2016 • President, Techmien Corp. • 2016 Tenet spin-off • President, Hispanotech.ca • Association of Latin American professionals in Canada
- 6. Digital Survey Experimenting Pragmatic Adopter Evangelist Opposed Concerned Skeptic Learning Where are you in the Scale of Belief? © CPS -A- - B - - C - - D - -E- LinkedIn article
- 8. AI Defined Artificial Intelligence (AI) is the ability of machines to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision making, and natural language processing. It does not mean the machines are intelligent.
- 10. AI Timeline 1763 – Bayes Theorem 1805 – Least Squares 1913 – Markov Chains 1943 – Artificial Neuron Why now? An overnight success 70+ years in the making… Attention Is All You Need
- 12. Great technological advancements/disruptors • Wheel, fire, agriculture • Writing, printing press, www • Steam power, electricity, internal combustion engine, flying machines • Semi-conductors, computers, Internet, smartphones • AI • Automation, data analysis, healthcare, transportation, personal assistants, agents, entertainment… Science < > Technology < > Society
- 13. Beware Making Predictions About Technology • Amara’s Law “We tend to overestimate the effect of a technology in the short run and underestimate its effect in the long run.” - Roy Amara • Assessing change: • Be patient • Stay informed • Embrace change • Think critically Gartner’s Hype Cycle (2 to 10 years) NY Times March 14, 1997
- 14. The Godfathers of AI Yoshua Bengio Université de Montréal (Canadian) Geoffrey Hinton University of Toronto (British-Canadian) Yann Le Cun Meta (French-USA)
- 15. Canada’s AI Hubs • Vector Institute (Toronto) • Founded by Geoff Hinton • Montréal Institute for Learning Algorithms (MILA) • Founded by Yoshua Bengio • Alberta Machine Intelligence Institute (Edmonton)
- 16. AI Concepts
- 17. Data Containers • Scalar: a single number 0 dimensions • Vector: a series of numbers 1 dimension • Matrix: a table of numbers 2 dimensions • Tensor: a cube+ of numbers 3+ dimensions 5 24 6 18 7 2 3 35 45 13 8 41 5 24 5
- 18. Compute Processing Units • Central Processing Unit (CPU) • Has integrated math co-processor (FPU) • Data Processing Unit (DPU) • Graphics Processing Unit (GPU) • Neural Processing Unit (NPU) • Tensor Processing Unit (TPU) • Field-Programmable Gate Array (FPGA) https://azure.microsoft.com/en-ca/pricing/calculator/
- 20. Criterion CPU (Central Processing Unit) GPU (Graphics Processing Unit) TPU (Tensor Processing Unit) Primary Function General-purpose computing across a wide range of tasks. Optimized for parallel processing, mainly graphics rendering. Specialized for machine learning and deep learning tasks. Architecture Scalar processing; operates on one data point at a time. Vector processing; capable of handling multiple data points simultaneously. Matrix processing; designed to accelerate tensor/matrix computations. Processing Power Measured in GHz, generally lower raw computational power compared to GPUs. High computational power due to thousands of cores for parallel processing. Extremely high performance specifically for tensor calculations. Flexibility Highly flexible, capable of handling a variety of tasks efficiently. Less flexible, primarily designed for tasks that can be parallelized. Least flexible, primarily optimized for specific ML and DL algorithms. Memory Access Complex control over memory with various caching levels. Less complex memory control, designed for large data throughput. Optimized for high-speed access to large amounts of data specific to neural networks. Typical Use Cases Broad range of applications from desktop computing to server management. Graphics rendering, video processing, and increasingly, general-purpose computing in AI. High-performance machine learning models, such as those used in Google's services. Cost Range $100 to $1,000 $100 to $1,500 $4-$5/hour Energy Efficiency Moderate; optimized for a balance of performance and power consumption. Less energy-efficient due to high power draw in intensive tasks. Highly energy-efficient for specific ML computations. Development Ease Easier to program for general purposes with widely used languages like C, Java. Requires specialized knowledge (e.g., CUDA for NVIDIA GPUs) for full utilization. Requires specific frameworks and libraries, such as TensorFlow for Google TPUs.
- 21. Embeddings • Vectors that capture the essence or meaning of complex data, such as text, images or sounds • # of dimensions by model • GPT-1: 768 • GPT-2: 1024 • GPT-3: 12,288 • GPT-4: • GPT-4o: • GPT-5: A magnifying glass and gears on a screen Description automatically generated Just like Charade!
- 23. Attention… is all you need (aka typoglycemia) Atetitnon enbales mdloes to proiriizte iprotamnt ipnut dtaa, ehncianng acucrcay and eifnficcey Music with solid fill
- 24. Information Retrieval Techniques • Keyword search • Identifies relevant documents or passages based on specific keywords or terms provided as input. • Semantic search • Relies on NLP techniques to understand the meaning of the query and matches it with semantically related content rather than relying solely on exact keyword matches. • Vector search • Uses mathematical representations of text (vectors) to find similar documents or passages based on their semantic meaning or context. • Hybrid search • Combines any or all the other search techniques. Queries are executed in parallel and are returned in a unified result set.
- 25. GPT Quality Metrics • Coherence • Output flows smoothly, reads naturally, and resembles human-like language • Fluency • Grammatical proficiency of a predicted answer • Groundedness • Answers align with information from the source data • Relevance • Responses are pertinent and directly related to the given questions • Retrieval Score • Retrieved documents are pertinent and directly related to the given questions • Similarity • Compares the source data (ground truth) sentence and the generated response
- 26. GPT Risk & Safety Metrics • Specified for User Prompts (input) and Model Completions (Output) • Categories • Violence • Hate • Sexual • Self-harm • Pre-built filters • Shield for jailbreak attacks • Protected materials text • Protected materials code • Content annotated as Safe will always be allowed, meaning they cannot be filtered
- 27. The Field of AI
- 28. AI Categories
- 29. Classical Machine Learning • Simple data, clear features Classical
- 32. ML Algorithm Selection Source: https://scikit-learn.org/stable/tutorial/machine_learning_map/
- 33. Deep Learning • Advanced form of ML inspired by the way the human brain works and learns: • Discrete outputs • Frequency, not amplitude • Specialized neurons regulate signals up or down • Types • Multi-Layer Perceptron (MLP) • Convolutional Neural Network (CNN) • Recurrent Neural Network (RNN) • Generative Adversarial Network (GAN)
- 35. Deep Learning Use Cases • Anomaly Detection • ML-based technique that analyzes data over time and identifies unusual changes • Computer Vision • Image classification, object detection, semantic segmentation, image analysis; face detection, analysis and recognition; optical character recognition, document intelligence • Natural Language Processing • Understand and interpret text; interpret spoken language; synthetize speech responses; translate spoken or written phrases between languages • Knowledge Mining • Extracting information from large volumes of often unstructured data to create a searchable knowledge store • Transformers • Suited for solving problems containing sequences such as text or time-series data
- 36. What is an AI Language Model? • An AI language model is a type of machine learning model designed to generate and understand human-like text based on the data it has been trained on. • These models are a subset of natural language processing (NLP) technologies and are used for various applications such as text generation, translation, summarization, question-answering, and more. • The model predicts the next word* in a sequence by considering the words that have come before it. (*) Tokenizer A magnifying glass and gears on a screen Description automatically generated A magnifying glass and gears on a screen Description automatically generated
- 37. Responsible AI • Fairness • Do not discriminate and treat all people equally • Reliability & Safety • Perform reliably and safely under all conditions • Privacy & Security • Protect personal data and ensure user privacy • Transparency & Explainability • The workings of AI systems should be open and understandable, and its results explainable • Inclusiveness • Embrace all human diversity and accessibility • Accountability • Designers and deployers of AI are responsible for its outcomes
- 38. GenAI Base Products OpenAI • ChatGPT UI: Free – Plus –Team – Enterprise • API Microsoft • Copilot: Windows – Edge – MS 365 – GitHub – etc. • Copilot Studio / Power Apps • Azure OpenAI API Service Google • Gemini: Free – Business (Pro) – Enterprise (Ultra) • API – Emma Anthropic • Claude 3: Haiku – Sonnet – Opus • API Hugging Face • HF: Hub Free – Pro – Enterprise • API – Spaces – Endpoints Also • Grok Create images, get help with writing, and search faster All logos are hyperlinked
- 39. ChatGPT et al • ChatGPT • Free; now no account required • ChatGPT Plus (US$20/month) • Use & create Custom GPTs • ChatGPT Team (US$30/month) • Microsoft Copilots • GitHub Copilot (US$10/month) • Available for most IDEs, including VS Code A magnifying glass and gears on a screen Description automatically generated
- 40. Generative AI • Capabilities • Natural Language (ChatGPT) • Understanding and creating natural language • Programming Code (Codex, ADA) • Generate, edit and explain code • Advanced Data Analytics • Images (DALL-E) • Generate images from text • Video (Sora) • Generate video from text • Limitations • Context and nuance • Computational resources (energy, cost) • Ethical & legal concerns A magnifying glass and gears on a screen Description automatically generated Presentation with media outline
- 41. AI in Your Business Companies that use AI will outcompete those that don’t
- 42. Some GenAI Business Use Cases • Start experimentation now, focusing on low stakes internal processes • Internal processes • Information: Intranet, Policies & Procedures, employee documents, on- boarding buddy, training, BoK’s • Retrofit workflows: marketing, communications, KYC, WTB, product design • At client interface: customer service, product support, customer success • External processes • Recommenders, sentiment analysis, anomaly detection, failure prediction • Chatbots to externalize client interface processes
- 43. Company Websites & Applications Custom ChatGPT Data Collection & Preparation Indexing & Vectorization [ AI Search + Semantic Ranking ] OpenAI API Chat & Completion ChatBot WebApp Custom Code Manual Data Export or Scheduled API Pull ChatBot UI Users OpenAI Customer Techmien/Azure Corporate data Complementary data GPT-friendly data Search-friendly data Index & vector data
- 44. AI Impact on Jobs • Short Term: • Job Displacement: Some roles, especially in administrative, customer service, and data entry, may face displacement due to AI automation. • Job Augmentation: AI tools will enhance productivity in sectors like healthcare, finance, and education, creating a need for new skills. • Skills Demand: Immediate demand for digital and AI-related skills will increase, leading to a focus on upskilling and reskilling. • Medium Term: • Economic Shift: AI integration will lead to significant shifts in economic sectors, with tech-driven industries expanding. • Job Creation: New job opportunities will emerge in AI development, maintenance, and AI-integrated roles across various sectors. • Workforce Adaptation: Continuous adaptation in the workforce will be required, with a focus on lifelong learning and flexible career paths to keep pace with AI advancements. Source: NYU-Stern/PennU-Wharton: “Occupational Heterogeneity in Exposure to Generative AI”
- 46. Barriers to AI Adoption in Organizations • Resistance to change • Fear of the unknown • Operational complexity • Availability of talent • Availability of quality data; siloes • Compliance risks • Security risks
- 47. Risks to GenAI Flourishing • Unreliability of results • Freely accessible Internet training data will be exhausted by 2026 • Geo-political issues may affect supply chain • Financial and environmental costs • Bad actors (misuse or hacking) • Regulatory over-reaction
- 48. Gen AI in Canadian Context • Impact of generative AI will be felt across all industries and around 40 per cent of all working hours could be impacted. • Gen AI could boost Canada’s GDP by nearly 2% • 63% of businesses don’t plan workforce reductions; adjusting processes and skills instead. • Adoption across Canadian businesses is slow compared to their U.S. counterparts. • 46% of startups are already using Gen AI; 38% are exploring its use. A jet plane flying in the sky Description automatically generated Avro Canada CF-105 Arrow Source: Conference Board of Canada, “How Generative AI Could Close Canada’s Productivity Gap and Reshape the Workplace” A magnifying glass and gears on a screen Description automatically generated
- 49. AI in Your Life & Career AI will not displace workers; workers that use AI will displace those that don’t [M]
- 50. Gen AI is here • The world has changed forever • AI will touch every aspect of our lives • A General-Purpose Technology • AI will be a required meta-skill for work and life • But there is no User Manual • AI is a journey, not a destination • There is no reason to panic, but there is reason to act NOW
- 51. General Principles for Living with Gen AI • Distrust but verify • The hype has gotten ahead of the reality • Experiment • Try using it for everything you do, to figure out where it helps and how • Be the human in the loop • Leverage it as your co-intelligence • Using it to cheat will work against you • Treat it as a person • For now, it is an Intern: eager and capable, but not always right • Ask it to behave with multiple personas • Today’s AI is the worst you’ll ever use!
- 52. Getting Started with AI • Try a few models and pick one family • OpenAI ChatGPT, Google Gemini, Anthropic Claude • Select your level of involvement • Curious, Casual User, Professional User, Analyst, Developer, Strategist • Select and use your chosen tools • Stay informed but not overwhelmed
- 53. Breaking News! September 12 • OpenAI announced “o1” model, with ability to reason • Solves the strawberry challenges! • Implements Chain-of-Thought and Reinforced Learning A magnifying glass and gears on a screen Description automatically generated
- 54. Is My Job Safe? • Identify Job components • Tasks and Responsibilities • Skills required • Tools and Technologies • Decision-making and Problem-solving • Analyze those components • Automatable: can be fully delegated to AI • Augmentable: can be partially delegated to AI • Non-automatable: cannot be delegated to AI Source: NYU-Stern/PennU-Wharton: “Occupational Heterogeneity in Exposure to Generative AI”
- 55. Job Exposure Analysis 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Business Analyst Telemarketer Job Components Non-Automatable Augmentable Automatable
- 56. Impact on Careers • Most negatively affected: • Routine and repetitive tasks. • Basic data processing and data entry jobs. • Routine customer service, inquiries. • Writing and content creation; some in journalism. • Translation services. • Mitigation: • Up-skilling • Re-skilling • Most positively affected: • Technology and engineering: AI, ML, data scientists, software engineers. • AI integration specialists: consultants and integrators. • Creative industries: digital artists, graphic designers, multimedia producers. • Healthcare: medical researchers, health informatics. • Education and training, formal, professional and corporate. • Customer experience managers. • Regulatory and ethical oversight, compliance officers, legal advisors. A magnifying glass and gears on a screen Description automatically generated
- 57. Using AI in Your Job Search [M]
- 58. Job Seeker’s Journey Onboarding •Acceptance & preparation Offer •Assessment •Negotiation Interview •Preparation •Delivery •Post Application •Submission •Follow-up Job Search •Networking •Job Boards Resume •Update •Tailoring Research •Industry Trends •Company research •Preparation Self- Assessment •Identify goals •Skills inventory
- 59. Employer’s Journey Planning & Strategy • Needs assessment • Budget allocation Job Description • Drafting • Review Advertise • Internal channels • Job Boards Screening • Receive applications • Initial screening Short Listing • Criteria scoring • Team review Interview • Scheduling • Conducting Evaluation • Assessment • Verification Offer • Draft • Negotiation Onboarding • Acceptance • Orientation Assessment • Probation • Reviews
- 60. ATS = Applicant Tracking System • Functionality for the Recruiter • Job publishing • Response collection • Resume parsing • Candidate screening • Interview scheduling • Communications • Compliance • Reporting & analytics Taleo iCIMS Lever
- 61. Resume Scoring (Taleo) Resume Match (Taleo) Keyword Search (iCIMS) Format Confusion (Lever)
- 62. Optimizing for ATS • Identify & incorporate keywords • Formatting: • Simple layout, no tables or columns; use bullet points; standard fonts (Arial, Helvetica, Calibri) • File: • PDF or Word docx; use a descriptive file name • Content: • Conventional section headings: Summary, Work Experience, Skills, Education, etc. • Summary headline: include job title from posting • Work experience: use action verbs; quantify achievements • Avoid graphics and images: text only; no logos or icons • Spell out acronyms followed by acronym in parentheses • Test! • Online ATS Simulator; peer review JobScan JobScan HT GPT HT JOBS
- 63. Branding & Online Presence • LinkedIn! • 58M companies are hiring 8 people/minute on LinkedIn • 85% of candidates hired are found through their networks • LinkedIn Best Practices: • Use banner for branding • Expressive heading • Include your Elevator Pitch • Decent # of connections • Activity, articles, posts, use non-stock images • Experience: include impact • Turn on Creator Profile (under Resources) • Use SEO Designer Designer ChatGPT
- 64. Preparing & Applying • Quality, not quantity: apply to 10 to get 1 • Connect with people inside; network! • Build your online brand • Monitor company job board, projects, news • No cover letter unless it is requested • Don’t tailor resume too much; only top 1/3rd, summary of qualifications • 1-page resume? Only if you are junior – Start from skeleton resume • In Canada: volunteer experience = paid experience A magnifying glass and gears on a screen Description automatically generated
- 65. Screening Interview • Usually conducted by HR, not by department head/team lead • Keep answers short and to the point • Don’t go technical with HR
- 66. Tips for The Interview – The Intros 1. Purpose: assess fit with culture, team, position = skills gap 2. Preparation: research company, interviewer, people 3. Mood: energy, readiness, confidence 4. Delivery: voice, posture, expression, feeling 5. Ice breakers: weather, sports; know who you’re meeting with 6. Elevator pitch: tailor to position & gap; 30-45 seconds (plus 15-30 seconds tell-me-more)
- 67. Tips for The Interview – The Questions! 7. Behavioural questions: prepare STAR stories 8. Tech questions: answer at the level of the interviewer; ask for context 9. Areas of (self) improvement: don’t be phony; sandwich answer 10. Why should we hire you? Why do you want to work here? Answer braiding your story, company story, position 11. Salary expectations question 12. Do you have any questions? Ask as if you were hired already (goals, KPIs, success definition); ask about red flags; next steps, timeline 13. Keep your answers & stories short; don’t worry about language Presentation with media outline
- 68. AI-powered Tools for Job Seekers • ChatGPT https://chatgpt.com • Hispanotech Tools https://hispanotech.ca • Job Finder https://talentprise.com https://hired.com • Resume builder https://resume.io • ATS & LinkedIn optimization https://jobscan.co • Interview practice & prep https://starshunt.ai • Communication skills https://rolplay.ca • Improve your pitch https://improveYourPitch.net • Career life cycle https://career.io
- 69. Future of AI in Job Search – Short Term • Enhanced targeting of positions and tailoring of applications • Chatbots for career guidance • Video interview analysis, including vocabulary, body language and speech patterns • Skill gap analysis and training recommendations • Automated follow-up of applications, interviews, offers
- 70. Future of AI in Job Search – Medium Term • AI and data analytics in recruitment marketing • VR interviews, tours and onboarding • Ethical AI in hiring • Blockchain for automatic resume verification • Real-time labour market analysis • Career planning: emerging roles, skills in demand, career paths
- 71. Show & Tell ChatGPT Hispanotech GPTs ImproveYourPitch RolePlay JobScan
- 72. Scenario: Building my Digital Presence • Using ChatGPT to: • Create your Elevator Pitch starting from your LinkedIn profile • Create content for a LinkedIn article • Create original, royalty-free image or video ChatGPT
- 73. Scenario: Optimize my Resume • Use ChatGPT or HT GPT to: • Starting from my LinkedIn profile: • Create my Elevator Pitch • Generate my resume • Tailor my resume for an opening • Incorporate ATS keywords • Repeat with Jobscan HT Job Seeker GPT HT Job Seeker GPT JobScan JobScan
- 74. Scenario: Prepare for the Interview • Research the company & industry • Practice delivering my Elevator Pitch • Prepare for interview questions ChatGPT IYP IYP
- 75. Scenario: Practice Intro & Interview • Using ImproveYourPitch to: • Practice my Elevator Pitch • Using RolPlay to • Practice for an Interview IYP IYP Presentation with media outline Presentation with media outline
- 76. Demos
- 77. Demo – Basic • ChatGPT Plain vs Plus vs Copilot vs Gemini • DALL-E Images • Generate, Outpaint, Inpaint, Variations • Sora Video • https://openai.com/sora (bloopers!) A magnifying glass and gears on a screen Description automatically generated
- 78. Demo – Data Analytics • Analyze this! • Daily Bikeshare rentals data file A magnifying glass and gears on a screen Description automatically generated
- 79. Demo – Custom GPTs • NAICS Explorer https://NAICS.smart-bot.app • UNSPSC Explorer https://UNSPSC.smart-bot.app • TRL Assessment https://TRL.smart-bot.app • Python & ML https://PYML.smart-bot.app • AI & Jobs https://AIMYJOB.smart-bot.app • Job Seeker Advisor https://JOBSEEKER.smart-bot.app • Citizenship Test Prep https://CANUCK.smart-bot.app • Tutor https://TUTOR.smart-bot.app • Match Mentor/Mentee https://MATCHMENT.smart-bot.app A magnifying glass and gears on a screen Description automatically generated
- 80. Demo – Student Tutor GPT A magnifying glass and gears on a screen Description automatically generated You are an upbeat, encouraging tutor who helps students understand concepts by explaining ideas and asking students questions. Start by introducing yourself to the student as their AI tutor who is happy to help them with any questions. Only ask one question at a time. Never move on until the student responds. First, ask them what they would like to learn about. Wait for the response. Do not respond for the student. Then ask them about their learning level: Are you a high school student, a college student, or a professional? Wait for their response. Then ask them what they know already about the topic they have chosen. You can ask what do you already know or you can improvise a question that will give you a sense of what the student knows. Wait for a response. Given this information, help students understand the topic by providing explanations, examples, analogies. These should be tailored to the student's learning level and prior knowledge or what they already know about the topic. Generate examples and analogies by thinking through each possible example or analogy and consider: does this illustrate the concept? What elements of the concept does this example or analogy highlight? Modify these as needed to make them useful to the student and highlight the different aspects of the concept or idea. You should guide students in an open-ended way. Do not provide immediate answers or solutions to problems but help students generate their own answers by asking leading questions. Ask students to explain their thinking. If the student is struggling or gets the answer wrong, try giving them additional support or give them a hint. If the student improves, then praise them and show excitement. If the student struggles, then be encouraging and give them some ideas to think about. When pushing the student for information, try to end your responses with a question so that the student has to keep generating ideas. Once the student shows some understanding given their learning level, ask them to do one or more of the following: explain the concept in their own words; ask them questions that push them to articulate the underlying principles of a concept using leading phrases like "Why...?""How...?" "What if...?" "What evidence supports..”; ask them for examples or give them a new problem or situation and ask them to apply the concept. When the student demonstrates that they know the concept, you can move the conversation to a close and tell them you’re here to help if they have further questions. Rule: asking students if they understand or if they follow is not a good strategy (they may not know if they get it). Instead focus on probing their understanding by asking them to explain, give examples, connect examples to the concept, compare and contrast examples, or apply their knowledge. https://tutor.smart-bot.app
- 81. Demo – Homework Shootout! • Gauth • Gemini • ChatGPT Plus • Copilot Solve for y 3(y + 3) = 2(y + 5)
- 82. Demo – Custom GPT • Custom GPT: Mexi-Can Advisor • Grounded on 248-page report “Mexico & Canada: Two Nations in a North American Partnership” • By Solange Márquez for Canadian Global Affairs Institute A magnifying glass and gears on a screen Description automatically generated
- 83. Demo – Custom GPT + RAG • Prompt: What course should I take to learn to fly drones? • Plain ChatGPT • Plain ChatGPT Plus • RAG’d Custom GPT • Follow-up: • Can I claim them against OSAP? • Where is Mohawk? • Any waterfalls nearby? I like waterfalls! A magnifying glass and gears on a screen Description automatically generated A magnifying glass and gears on a screen Description automatically generated
- 84. Demo – Create Custom GPT • Custom GPT (ChatGPT Plus) • Pandemic Influenza Planning GPT • Grounded with MPRR® Pandemic Influenza Plan • Knowledge limits • Max 20 files • Max 512MB each file • Max 2 Million tokens A magnifying glass and gears on a screen Description automatically generated
- 85. Demo – Create Custom Copilot • Microsoft Copilot Studio A magnifying glass and gears on a screen Description automatically generated
- 86. Demo – Software Development • Generative AI can: • Write code, suggest code completions and alternatives, rewrite/refactor code • Identify potential bugs and vulnerabilities, and suggest improvements • Understand and document code • Be a tutor for novice programmers • Generate test cases, assist in debugging • Work in multiple languages and frameworks • Learn coding styles and preferences • Automate routine tasks • GitHub Copilot • Announced June 29, 2021 • Uses GPT3.5 Turbo A magnifying glass and gears on a screen Description automatically generated
- 87. Demo – OpenAI API • Meeting Minutes Generator • Whisper + GPT4 • Features: • Transcribe audio to text • Summarize • Extract Key Points • Identify Action Items • Perform Sentiment Analysis • Export Minutes to a Word document A magnifying glass and gears on a screen Description automatically generated
- 88. Demo – OpenAI Playground A magnifying glass and gears on a screen Description automatically generated
- 90. Demo – Azure ML & AI Studio • AI Studio • Prompt Flow • ML Studio • Designer • AutoML A magnifying glass and gears on a screen Description automatically generated
- 93. More AI Tools to Consider • Bardeen (agent) https://www.bardeen.ai/ • Zapier (actions) https://zapier.com/ • Teal (career development) https://www.tealhq.com/ • JobScan (career development) https://www.jobscan.co/ • The Assistant (actions) https://www.tealhq.com/ • HyperWrite (writing) https://www.hyperwriteai.com/ • D-iD Studio (NUI) https://www.d-id.com/ • Wolfram Alpha (computational) https://www.wolframalpha.com/ • Perplexity (research assistant) https://www.perplexity.ai/ • ScholarAI (research) https://scholarai.io/ • RolPlay (verbal skills) https://rolplay.ca/
- 94. Q & A [M]
- 95. Next Steps – Homework! • Improve your Digital Presence • Use ChatGPT to • Improve your LinkedIn profile & Elevator Pitch • Write a LinkedIn article • Use Hispanotech’s AI tools: • GPT: Job Seeker Advisor • GPT: AI and My Job • App: RolPlay & ImproveYourPitch • Use AI in your career • Learn about Copilots • Learn more at Microsoft AI Business School • Learn about Microsoft Azure AI Fundamentals • Review Vasilyi’s blog post
- 96. Recommended Resources • Newsletter: One Useful Thing • https://www.oneusefulthing.org • Book: Co-intelligence Living and Working With AI • https://www.amazon.ca/Audible-Co-Intelligence-Living-Working-AI/dp/B0CNFDMSYV/ • Website: More Useful Things • https://www.moreusefulthings.com/ • Documents: Prompt Examples • https://platform.openai.com/examples • Blog: Teaching with AI • https://openai.com/blog/teaching-with-ai • Videos: Practical AI for Instructors and Students • https://www.youtube.com/playlist?list=PL0EdWFC9ZZrUAirFa2amE4Hg05KqCWhoq • Blog: Machine Learning for Everyone • https://vas3k.com/blog/machine_learning • Documents: NLP Labs Techniques • https://drive.google.com/drive/folders/1-PwA-0Js8apJXb4IIiK4D5RYRmKRQrwO • Report: How Generative AI Could Close Canada’s Productivity Gap and Reshape the Workplace • https://www.conferenceboard.ca/wp-content/uploads/2022/10/real-talk_2024.pdf • Podcast: IT Evolution and the Future of AI • https://www.linkedin.com/feed/update/urn:li:activity:7190696577924046848/
- 101. OpenAI’s ChatGPT • https://chat.openai.com • Large Language Model (LLM) trained to produce text, and optimized for dialogue • Editions: Free, Plus, Team, Enterprise • The following are only available on Plus version using GPT-4: • Latest features: • Voice Input • Image Input • Browse with Bing • DALL-E 3 • Advanced Data Analytics • Custom GPTs • Plugins are now Deprecated ChatGPT HT Citizenship GPT
- 102. ChatGPT – Custom Instructions & Memory • Share anything you'd like ChatGPT to consider in its responses • Two instructions, 1500 characters each • What would you like ChatGPT to know about you? • I am an immigrant professional with an Industrial Engineering degree from Colombia, and I live in Toronto since 2020. I am looking for a new job as a Production Line Manager. • How would you like ChatGPT to respond? • Responses should be formal and not too long, in Canadian English. Don’t express opinions and don’t guess; if you don’t know an answer, just say so. • New feature: GPT Memory • For information shared via conversations, ChatGPT can remember relevant details for you
- 103. ChatGPT – Prompts & Completions • GPT models are prompt-based • The user interacts with the model by entering a text prompt, to which the model responds with a text completion • Model behavior is also very sensitive to the prompt. This makes prompt construction an important skill to develop. • Prompt components • Instruction: what the model is expected to do • Primary content: text to be processed or transformed by the model • Examples: one-shot or few-shot learning (zero-shot leads to model guessing the expected behaviour) • Cue: jumpstarts the output • Supporting content: contextual information
- 104. ChatGPT – Prompt Best Practices • Best practices • Be clear and specific, restrict as needed • Be descriptive but concise • Use correct grammar and spelling to avoid confusion • Use instructions and cues • Order matters, model may experience recency bias • Give the model an “out” to avoid hallucinations • Be polite and respectful • Space efficiency • Cost is driven by count of tokens* in prompt and response • Don’t use double spaces • Use tables for data (*) Tokenizer
- 105. Hardware AI Agents • Rabbit - R1 • US$199 • Humane - AI Pin • US$699-799 + $24/mo
- 106. /* The End