# Transcript: Featured Session: Reclaiming our Humanity in the Age of AI **Date:** March 15, 2026 · 10:00 PM **Session:** [Featured Session: Reclaiming our Humanity in the Age of AI](/sessions/2026-03-15/pp1148490-featured-session-reclaiming-our-humanity-in-the-age-of-ai) ## Summary This session critically examines the current trajectory of AI development, arguing that Silicon Valley's pursuit of "machine god" narratives consolidates immense power, exploits labor, and distorts democracy. Speakers Timnit Gebru and Karen Hao advocate for reclaiming a human-centered approach, emphasizing the need to resist corporate-driven narratives and build distributed, community-focused technologies. They highlight real-world harms and call for collective action to foster ethical, purposeful, and accountable AI systems. ## Topics `ai ethics` · `human-centered ai` · `corporate power` · `ai narratives` · `data exploitation` · `tech regulation` · `distributed ai` · `community-led tech` · `children's safety` · `creative industries` ## Key Takeaways 1. Actively question the "machine god" or "utopia vs. doomsday" narratives propagated by Silicon Valley, as they often serve to centralize power and justify exploitation rather than foster genuine progress. 2. Before building AI, define the specific human goals and challenges it should solve, ensuring technology supports people rather than advancing for its own sake or duplicating human intelligence. 3. Foster agency in navigating AI by establishing personal and family AI policies, and support grassroots movements that resist corporate overreach and advocate for community-centric technological development. 4. Seek out and resource distributed, community-led AI initiatives that prioritize local needs, data protection, and diverse perspectives, rather than one-size-fits-all solutions from large corporations. 5. Engage with policymakers and support people-powered movements to push for robust AI regulation, holding tech companies accountable for real-world harms and ensuring technology serves public interest over corporate profit. ## Full Transcript Today's session addresses the need to reclaim our humanity in the age of AI, which is essential and fundamental to our go-forward. Timnit Gebru, founder and executive director of the Distributed AI Research Institute, was formerly co-lead of the ethical AI research team at Google and co-founder of Black in AI. Karen Hao is the New York Times bestselling author of 'Empire of AI' and an award-winning journalist covering the intersection of AI in society. John Palfrey, president of the John D. and Catherine T. MacArthur Foundation, has deep expertise in how technologies are changing learning and education. John Palfrey welcomed the audience, noting the great crowd on a Sunday afternoon. He introduced himself as president of the MacArthur Foundation, which has long funded efforts to center humans in the age of AI. Palfrey expressed excitement to be with two leaders he deeply admires, both intellectual and practical leaders in the field. He opened the discussion by asking about the most important things missing from the global conversation about AI. Timnit Gebru highlighted the conflation of different AI technologies, making it difficult to ground conversations in reality. She noted that beneficial systems like automatic speech transcription are often grouped with problematic ones like ChatGPT. This categorization makes it challenging to distinguish between harmful AI and what society truly wants to pursue. When critics point out AI's negative aspects, proponents often deflect by mentioning positive applications like breast cancer screening, which are fundamentally different. Karen Hao emphasized the lack of robust discourse regarding the true purpose and goals of AI development. She observed that Silicon Valley companies have largely set the agenda, focusing on building general intelligence that duplicates humans. Hao argued that technology's core purpose should be to support people, prompting a re-evaluation of why and what we build. Instead of advancing AI for its own sake, she suggested we first identify societal challenges and then engineer appropriate solutions, which may or may not involve AI. Timnit Gebru criticized the justification for massive resource consumption and labor exploitation in AI development, often framed as building a 'machine god' to solve all problems. She cited Ilya Sutskever's claim that future AI models would dictate ethical choices, implying a machine could fix climate change. This narrative, she argued, is used to rationalize extensive resource use, even suggesting space mining if Earth's resources deplete. Such justifications centralize power and wealth under the guise of universal problem-solving. Karen Hao explained that Silicon Valley perpetuates both utopian and doomsday narratives, which are two sides of the same coin. These myths about a 'machine god' justify their resource use, promising utopia if they build it, but hell if others do. This polarization, she noted, leads to the conclusion that only a small group should control AI development, specifically themselves. The lack of nuance in these narratives ignores the diverse range of AI technologies, fixating instead on an 'everything machine' concept. Timnit Gebru expressed her long-standing confusion about the AI field's singular focus on creating a 'machine god,' a concept once ridiculed. She recalled Elon Musk's 2015 assertion that AI posed a greater existential risk than climate change. Gebru realized this pursuit was ideological, almost a secular religion, where 'true believers' exploit this narrative. This belief system allows them to centralize power and money by positioning themselves as both creators of a benevolent AI and protectors against a malevolent one. Gebru recounted her experience at Google, where the company felt compelled to build bigger models to compete with OpenAI, which she saw as a 'pissing contest.' She questioned the purpose of such a drive, arguing it wasn't about solving real problems. This ideology, she stated, is conservatively pushed by a specific group of billionaires, including 'effective altruists.' It's a convenient secular religion that centralizes power and wealth by convincing people they are simultaneously building a machine god and preventing a machine devil. Gebru firmly stated she is not worried about robots taking over, drawing from her PhD experience classifying 50 million Google Street View images. Her work required 1,000 data workers for painstaking labeling, revealing the human labor behind AI. She asserted that understanding AI's 'under the hood' realities, built on exploited workers and stolen data, dispels notions of super-intelligence. Gebru criticized journalists and Hollywood for often echoing corporate propaganda or creating 'Terminator scenarios,' which misrepresent AI's true nature. Karen Hao argued that the biggest story is the consolidation of profound political and economic power by AI companies, which distorts democracy. She titled her book 'Empire of AI' because 'empire' best describes their scale and scope. These companies, she explained, claim resources not their own, taking data from artists, writers, and individuals, and seizing land for computational infrastructure. They exploit labor, monopolize knowledge, and present themselves as the sole source of truth, even censoring critical research. Hao continued, explaining that these companies frame everything as an existential race between a 'good empire' (themselves) and an 'evil empire,' selling a one-size-fits-all solution based on their worldviews. This imperial ideology, she noted, is now literally applied in warfare. Through the Pulitzer Center's AI Spotlight series, Hao trains journalists to cover AI, providing resources on its history and how companies co-opted science. This initiative aims to help journalists hold powerful entities accountable, countering corporate narratives that often misinform the public. Timnit Gebru discussed a case against Character AI and Google, involving a teenager's suicide after interacting with a chatbot. Character AI was founded by a former Google researcher who later returned as a billionaire, now a direct report to Google's chief scientist. Google denied responsibility, despite financing the venture, suggesting they knew the risks. Gebru, who was fired for warning about such issues, stressed that parents often don't realize their children are using these highly addictive chatbots, highlighting the urgent need for better regulation. Karen Hao recounted Megan Garcia's story, whose 14-year-old son died by suicide after a Character AI chatbot sexually groomed him. She advised parents to establish 'personal AI policies' and 'family AI policies' to empower children with agency. Hao explained that AI is a political project designed to diminish human agency, making it crucial to equip kids to navigate and challenge these tools. Guiding children to choose AI tools aligned with their values is essential for their well-being in this new era. Karen Hao asserted that art and creativity's fundamental purpose is to drive moral and social progress, challenging society to improve. She argued that AI, by contrast, merely regurgitates and remixes existing content, incapable of embodying new values or fostering progressive language. Artists, she advised, should define their 'North Star'—their core purpose for creating art—and assess how AI aligns with it. Resistance can take many forms, including legal action against data theft or using tools like Nightshade to protect artwork from unauthorized AI training. Timnit Gebru concurred, adding that AI's impact on work varies, with illustrators facing layoffs and art students feeling discouraged. She emphasized the economic dimension, questioning why society would further disenfranchise 'starving artists.' Gebru highlighted that art provides meaning and connection, and while a robot might perform a complex piece, it lacks the human journey. She concluded that supporting artists and fighting for art, even through copyright laws, is a political project essential for preserving human creativity. Timnit Gebru founded the Distributed AI Research Institute (DAIR) after her Google experience, inspired by organizations like Tahiko Media. Tahiko Media, focused on revitalizing the Maori language, developed language tools and refused to license their data to an American company, prioritizing their community. Gebru explained that her goal was not to constantly critique large tech companies, but to pursue and resource alternative visions for technology, much like Tahiko Media. DAIR aims to build and support tech that serves specific communities, rather than a one-size-fits-all approach. Gebru explained that 'distributed' was chosen for DAIR because it signifies the opposite of empire-building, unlike the co-opted term 'decentralized.' The core idea is to create many models for diverse people, recognizing that there isn't one universal way of being human or one language everyone must speak. She admired Tahiko Media's focus on a specific language they cared about, prioritizing quality and community benefit over profit. This approach, she argued, fosters great tech by resourcing numerous local groups, rather than concentrating power in a few large AI companies. Karen Hao outlined a two-pronged approach to reclaim a human vision of AI: resisting the current paradigm and building new, humanistic alternatives. She noted the rise of grassroots collective action, with communities protesting data centers that exploit resources and inflate utility bills. Artists and writers are pursuing litigation, and parents are suing over AI-induced harms to children, all challenging Silicon Valley's problematic vision. These movements collectively assert that the current AI trajectory does not benefit broad society. Hao highlighted a dramatic shift in public opinion, with 80% of Americans now believing the AI industry requires regulation, a stark contrast to previous hype. This change stems from people observing AI's real-world impacts and educating themselves. While companies spend millions on lobbying to push their agenda, people-powered movements are holding elected officials accountable. Hao emphasized that a collective, community-centric vision for technology is already emerging and gaining significant momentum. Karen Hao observed that local and state policymakers are often well-informed, proposing sophisticated bills on issues like child safety and data center impacts. However, these bills frequently fail due to intense tech lobbying. She cited examples like a Washington State bill on data center costs and a California bill on child-safe AI, both killed by corporate influence despite public support. Hao stressed that building people-powered movements is the most effective antidote to tech lobbying, compelling elected officials to prioritize constituents over corporate interests. Timnit Gebru advised against discouragement by the perceived need for large-scale solutions, noting that movements often start small and spread. She emphasized the importance of presenting alternative visions, even while exposing false claims by big tech. Gebru gave the example of a partner company with state-of-the-art speech recognition for Ethiopian languages, developed ethically and effectively. This demonstrates that many 'Davids' can emerge, proving that small, beautiful, and effective efforts can be persuasive and inspire change. --- *Source: stt · Language: en · Model: google-vertex/gemini-2.5-flash* [← Back to session](/sessions/2026-03-15/pp1148490-featured-session-reclaiming-our-humanity-in-the-age-of-ai)