Knowledge Vault 4 /34 - AI For Good 2019
Interview
Héléne Molinier
< Resume Image >
Link to IA4Good VideoView Youtube Video

Concept Graph & Resume using Claude 3 Opus | Chat GPT4o | Llama 3:

graph LR classDef unWomen fill:#d4f9d4, font-weight:bold, font-size:14px classDef tech fill:#f9d4d4, font-weight:bold, font-size:14px classDef aiAssistant fill:#d4d4f9, font-weight:bold, font-size:14px classDef genderBias fill:#f9f9d4, font-weight:bold, font-size:14px classDef principles fill:#f9d4f9, font-weight:bold, font-size:14px A[Interview] --> B[UN Women champions
gender equality in AI. 1] A --> C[Tech seen as neutral,
but impacts differ. 2] A --> D[Female-voiced AI assistants,
male designers. 3] D --> E[AI assistants: men get Viagra,
women not contraception. 4] D --> F[Amazons AI resume
screening favored men. 5] A --> G[Girls equal ability, norms
steer from AI. 6] G --> H[Gender inequity stems from
social norms. 7] A --> I[UN Women created
Gender Innovation Principles. 8] I --> J[Principles: high-level commitment,
gender in design. 9] I --> K[More women in AI,
staff gender training. 10] A --> L[Summit message: cross-sector
efforts needed. 11] A --> M[Summit raised awareness
of gender lens. 12] A --> N[Social sciences, tech must
discuss gender. 13] class A,B,F,G,H,L,M,N unWomen class C,D,E tech class D,E aiAssistant class I,J,K principles class G,H genderBias

Resume:

1.- UN Women champions women's rights, gender equality, and women's empowerment. It's important for them to address the gender imbalance in AI.

2.- Technology is often seen as gender-neutral, but AI solutions frequently impact women and men differently, necessitating UN Women's involvement.

3.- Examples: People prefer giving orders to female-voiced AI assistants. The assistants' responses to sexual questions, designed mostly by men, are sometimes shocking.

4.- AI assistants provide different answers to men and women, e.g. giving men directions to find Viagra but not helping women find contraception.

5.- Amazon's AI resume screening tool favored men, setting aside resumes mentioning the word "women", demonstrating bias in AI systems.

6.- While girls and boys have equal ability based on test scores, social norms and lack of opportunities steer girls away from AI careers.

7.- Gender inequity in AI stems from social/cultural norms and stereotypes. Changes are needed in government, business and education to address this.

8.- UN Women created Gender Innovation Principles with 30 partners to provide concrete solutions organizations can adopt to improve gender equity in AI.

9.- The principles include: 1) Secure high-level commitment from senior management to prioritize gender. 2) Integrate gender considerations in the AI design stage.

10.- Additional principles: 3) Include more women on AI teams. 4) Train all staff, especially technical staff, on gender issues and unconscious biases.

11.- UN Women's message at the summit targeted researchers, the private sector, and governments, as the issue requires cross-sector efforts to fix.

12.- The interviewee felt the summit made progress in raising awareness, with attendees saying they hadn't previously considered these issues through a gender lens.

13.- It's important for the social sciences and technology to have more discussions together to address gender inequity in AI.

Knowledge Vault built byDavid Vivancos 2024