Meta announced the formation of an AI advisory council composed entirely of white men on Wednesday. This move highlights the ongoing issue of exclusion and underrepresentation of women and people of color in the AI industry, despite their significant contributions and qualifications.
Unlike Meta’s more diverse board of directors and Oversight Board, this new AI advisory council lacks gender and racial representation. This council, not elected by shareholders and without fiduciary duty, aims to provide insights and recommendations on technology advancements and strategic growth opportunities, meeting periodically.
Composition and Expertise of the Advisory Council
The council consists entirely of businesspeople and entrepreneurs, without the inclusion of ethicists or academic researchers. While executives from companies like Stripe, Shopify, and Microsoft might have experience bringing products to market, AI poses unique risks, especially for marginalized groups.
Sarah Myers West, managing director at the AI Now Institute, emphasized the need to critically examine AI-producing institutions to ensure they serve the public’s needs. She highlighted that AI errors disproportionately affect communities that have historically faced discrimination, underscoring the importance of a higher standard in AI development.
Women are significantly more vulnerable to AI’s negative impacts. A 2019 study by Sensity AI revealed that 96% of AI-generated deepfake videos were nonconsensual and sexually explicit. High-profile cases, such as nonconsensual deepfakes of Taylor Swift, illustrate the ongoing threat to women from AI technology.
Ordinary individuals face significant challenges in combating AI-generated harm, unlike high-profile victims like Swift. Reports indicate middle and high school students creating explicit deepfakes of classmates, facilitated by easily accessible apps. Instances of apps like Perky AI, which Meta initially failed to detect, further demonstrate the systemic issues in moderating harmful AI content.
The Importance of Inclusive AI Development
The exclusion of women and people of color from AI innovation has led to disastrous outcomes. Historical examples, such as the exclusion of women from clinical trials until the 1970s, show the long-term consequences of biased research. Similarly, AI technologies have shown biases against Black people, such as self-driving cars being more likely to hit them due to sensor limitations.
AI systems often perpetuate existing biases in areas like employment, housing, and criminal justice. For example, voice assistants struggle with diverse accents, and facial recognition systems disproportionately flag Black individuals. The development of AI reflects existing power structures and fails to address systemic biases.
The rapid development of generative AI could exacerbate existing inequalities. According to McKinsey, AI could automate nearly half of all jobs that do not require a four-year degree and pay over $42,000 annually, roles in which minority workers are overrepresented.
There is significant concern about an all-white male team at a leading tech company advising on AI products for a diverse population. Developing safe and inclusive AI requires a comprehensive understanding of societal issues and diverse perspectives. While Meta’s advisory board may fall short, it leaves room for other startups to rise and address these critical gaps.
See also: Orca AI Secures $23M To Advance Autonomous Shipping Technology
