The Responsible AI Game by JL Marechaux

Embark on a Responsible AI Quest! Gain valuable insights into the ethical considerations surrounding AI and learn how to mitigate potential harms. And Join the game and compete for the chance to be crowned the 'Guardian of Ethical AI'

What to Expect:

* Impactful Learning: Engage with key AI concepts. No prior AI knowledge is required!

* Interactive Gameplay: Play against fellow questers to tackle challenges in a fun and engaging ethical AI game.

* Ethical Exploration: Inspect real-world dilemmas and learn how to apply ethical frameworks to AI development and deployment.

https://youtu.be/ytwPEnqY8Tc

DevFest Montreal 2024

Transcript

1. The Responsible AI Game Play. Learn. Level-up your AI ethics. Montreal JL Marechaux Google
2. Google Solutions & Thought Leadership (Data Science, ML & GenAI) JL Maréchaux • Expertise in Software Engineering & Cloud Computing • Advanced Analytics & Applied Artificial Intelligence • Data Science & AI/ML for Digital Marketing AI projects with companies in North America (“custom” AI models) • (20%) Google AI Principles Pioneers
3. Ethical AI is everyone's responsibility As data scientists, citizens, ML practitioners, students and AI product users, we must educate ourselves and others on the benefits and the potential risks of AI. We must advocate for an AI that is good for individuals, society and the environment.
4. This presentation reuses existing images. It does not include new AI-generated visuals. #WeAreResponsibleAI
5. Power & complexity The need for Responsible AI time AI & Tech Law & regulations Governance Gap
6. Power & complexity The need for Responsible AI time AI & Tech Law & regulations Ethics & Responsible AI
7. Socially beneficial Privacy Safety & Security Interpretability Fairness AI governance Ethics Explainability Sociotechnical Robustness Eco-Responsible Transparency Human-controlled Accountability Trustworthy Responsible AI: A set of principles and practices for developing and deploying artificial intelligence (AI)
8. 1 Be socially beneficial Thoughtfully evaluate benefits for the community. Avoid creating or reinforcing unfair bias Seek to avoid unjust impacts on people. 2 Be built and tested for safety Strong safety & security practices to avoid unintended harm. 3 Be accountable to people Appropriate human direction and control. 4 Incorporate privacy design principles Transparency, control, and consent data. 5 Uphold high standards of scientific excellence Based on scientifically rigorous and multidisciplinary approaches. Google AI Principles 6 7 Be made available for uses that accord with these principles Limit potentially harmful or abusive applications.
9. Responsible AI toolkits: Ethical frameworks and guidelines for the development and deployment of AI Know your data (beta) Data Cards Fairness indicator What-If tool TF Privacy …. Design Safety Evaluation Safeguards 20+ frameworks 15+ Tools and designs Explainable AI Model Cards Predictive AI & Generative AI Google Cloud Tools and practices Vertex AI BigQuery
10. Let's play a game
11. Video (First 45 s) https://www.youtube.com/watch?v=S75NcDqbPbI
12. Actuator Display zone Biometric cameras Body sensors TuringTee, the AI-Powered T-Shirt (prototype) Geolocation (GPS) Mics
13. AI 1 2 3 4 5 6 7 Use Responsible AI principles to guide the development of the TuringTee prototype You "Responsible AI Advocate"
14. Use your own device: Mobile, tablet or laptop Play alone, or with others Find the best "Responsible AI" answers Quick answers get more points The "Guardian of Ethical AI" wins a prize Game rules
15. Responsible AI Game (30+ min)
16. Last word
17. Look for Responsible AI frameworks and guidelines near you Montréal declaration for a responsible development of artificial intelligence https://montrealdeclaration-responsibleai.com Google Responsible AI https://ai.google/responsibility/principles Artificial Intelligence and Data Act (AIDA) https://ised-isde.canada.ca/site/innovation-better-canada/en/artificial-intelligence-and-data-act-aida-companion-document Ethics of AI Responsible AI https://www.unesco.org/en/artificial-intelligence/recommendation-ethics
18. Shameless plug Beyond the Hype: When Generative AI Isn’t Always the Answer Why predictive AI might still be your best bet Published in Towards Data Science, October 2024 https://medium.com/towards-data-science/beyond-the-hype-when-generative-ai-isnt-always-the-answer-3ddbb19ce70f
19. Q & A
20. Merci | Thank You JL Maréchaux Google November 2024