The End Of Knowledge - Vault 4/73 - AI For Good - 2022

graph LR classDef event fill:#f9d4d4, font-weight:bold, font-size:14px classDef ethics fill:#d4f9d4, font-weight:bold, font-size:14px classDef speaker fill:#d4d4f9, font-weight:bold, font-size:14px classDef privacy fill:#f9f9d4, font-weight:bold, font-size:14px classDef fairness fill:#f9d4f9, font-weight:bold, font-size:14px classDef governance fill:#d4f9f9, font-weight:bold, font-size:14px A[Responsible AI in
practice] --> B[Organized by ITU
with 40 UN organizations. 1] A --> C[Identify AI for
UN goals impact. 2] B --> D[Use live video wall
for engagement. 3] A --> E[UNESCO intro:
responsible AI importance. 4] A --> F[Ethical approach:
human rights, dignity focus. 5] F --> G[UNESCO AI ethics:
rights, inclusiveness, peace. 6] G --> H[UNESCO ethics adopted
by 193 states. 7] A --> I[Dr. Nashlie Sephus on
responsible AI. 8] A --> J[Respect privacy,
fairness, explainability. 9] J --> K[AI systems must
respect values. 10] J --> L[Gender classifiers show
need for bias mitigation. 11] J --> M[Over 50% US faces
in datasets. 12] J --> N[AWS halts law enforcement
facial recognition. 13] A --> O[AI spending $204 billion
by 2025. 14] A --> P[Diverse AI leadership
crucial. 15] A --> Q[Responsible AI respects
human rights. 16] A --> R[Operationalizing responsible
AI challenging. 17] A --> S[AI systems need
clear rationales. 18] A --> T[Robust against attacks,
transparent operations. 19] A --> U[Governance must enforce
responsible AI. 20] U --> V[Ethical data collection
vital. 21] U --> W[Promote AI ethics
education early. 22] U --> X[Governments regulate
AI use. 23] U --> Y[Businesses accountable for
AI practices. 24] A --> Z[Balance fairness, privacy,
performance. 25] A --> AA[Equitable opportunities
for all demographics. 26] A --> AB[Address intersectional
fairness in AI. 27] A --> AC[Diverse feedback improves
AI fairness. 28] A --> AD[Monitor AI systems
continuously. 29] A --> AE[Raise public awareness
on AI. 30] class A,B,C,D,E,F,G,H,I,J,K,L,M,N,O,P,Q,R,S,T,U,V,W,X,Y,Z,AA,AB,AC,AD,AE event class F,G,H ethics class I speaker class J,K,L,M,N privacy class Z,AA,AB fairness class U,V,W,X,Y governance

Resume:

1.- Event Organization The event was organized by the ITU in partnership with 40 UN sister organizations, co-convened with Switzerland, focusing on practical AI applications.

2.- Event Goal Aim to identify AI applications to advance UN sustainable development goals and scale solutions for global impact.

3.- Audience Engagement Participants encouraged to use the live video wall for questions, comments, and engaging discussions with panelists and experts.

4.- Session Introduction Introduction by interim director of UNESCO’s Social and Human Sciences sector, emphasizing responsible AI’s importance.

5.- Ethical Approach Emphasis on the need for an ethical approach to AI, focusing on human rights and dignity.

6.- UNESCO’s Core Values UNESCO's AI ethics recommendation includes respect for human rights, environmental flourishing, inclusiveness, and peaceful societies.

7.- Recommendation Adoption UNESCO's AI ethics recommendation adopted by 193 member states in 2021, highlighting global commitment to responsible AI.

8.- Keynote Speaker Dr. Nashlie Sephus, expert in machine learning and algorithmic bias, shared insights on responsible AI and industry practices.

9.- Privacy and Fairness Responsible AI should respect privacy, fairness, explainability, robustness, transparency, and governance.

10.- Ethical AI Systems AI systems must respect values and address privacy, fairness, and transparency to serve humanity effectively.

11.- Bias in AI Example of gender classifiers potentially discriminating against non-binary people, highlighting the need for bias mitigation.

12.- Privacy Concerns Over 50% of people’s faces in the US are included in datasets without their knowledge, raising privacy issues.

13.- Facial Recognition AWS placed a moratorium on selling facial recognition tech to law enforcement due to privacy and bias concerns.

14.- AI’s Economic Impact Global spending on AI projected to reach $204 billion by 2025, with significant organizational transformation expected.

15.- Diverse Leadership Nurturing diverse leaders in AI is crucial for ensuring responsible and inclusive technological development.

16.- Human Rights Respect Responsible AI must be innovative, trustworthy, and respect human rights and democratic values.

17.- Operational Challenges Organizations struggle with operationalizing responsible AI despite recognizing its transformative potential.

18.- AI Model Explainability Ensuring AI systems offer clear rationales for their decisions is crucial for accountability and trust.

19.- Robustness and Transparency AI systems should be robust against adversarial attacks and transparent to users about their operations.

20.- Governance in AI Governance structures must enforce responsible AI practices across all stakeholders involved in the AI lifecycle.

21.- Ethical Data Collection Collecting data ethically, ensuring consent, and maintaining confidentiality are vital for responsible AI development.

22.- AI Education Promoting AI and ethics education from an early age can help build a more informed and responsible AI workforce.

23.- Government’s Role Governments should ensure appropriate AI use and develop risk-based regulations with input from various stakeholders.

24.- Industry Accountability Businesses must be accountable for responsible AI practices, with documentation and external assessments to ensure compliance.

25.- Trade-offs in AI Balancing trade-offs in fairness, privacy, and performance is essential for developing responsible AI solutions.

26.- Equitable Opportunities AI systems should provide equitable opportunities, considering demographic disparities and marginalized groups.

27.- Intersectional Fairness Addressing intersectional fairness is critical to ensure AI systems do not disproportionately harm specific subgroups.

28.- Stakeholder Feedback Incorporating feedback from diverse stakeholders helps refine and improve AI systems’ fairness and effectiveness.

29.- Continuous Monitoring Ongoing monitoring and evaluation of AI systems are necessary to maintain their performance and fairness over time.

30.- Public Awareness Raising public awareness about AI’s benefits and risks is essential for fostering trust and responsible adoption of technology.

Knowledge Vault built byDavid Vivancos 2024