Partnership on AI

I am pleased to join such a diverse group of thought leaders on the Partnership on AI Labor and Economy Steering Committee. At the AFL-CIO Tech Institute, our mission is to ensure that technological change creates real economic opportunity and centers the knowledge and interests of working people. Joining conversations with leaders across labor, industry, academia, and civil society is vital to that effort. As we face this watershed moment, when new technologies will shape the future of millions of jobs, the Tech Institute is committed to advancing solutions that protect workers' rights and support their lives and livelihoods. I look forward to the work ahead!

Today's announcement from Partnership on AI marks a meaningful step in taking AI's economic consequences seriously. PAI has just launched its Labor and Economy Steering Committee — a new expert group bringing together labour leaders, economists, civil society organisations and policymakers to develop concrete recommendations on AI's economic consequences. Windfall's scenario analysis will be at the heart of that work. The committee's membership speaks to the seriousness of the moment: Anton Korinek, Harvard University's Dani Rodrik, AFL-CIO Tech Institute, The World Bank Group, Economic Policy Institute Google DeepMind and OpenAI's chief economist among others. These are not people who gather to produce reports that sit on shelves. What Windfall brings to this collaboration is structured frameworks that allow serious people to stress-test assumptions about what AI disruption actually looks like, what policy levers are available, and what happens when governments act before the crisis arrives rather than after. Our London workshop in February, which brought together over seventy economists, policymakers and AI researchers, showed what that kind of structured preparation can surface: not reassurance, but clarity about what preparation actually requires. The decisions being made now about how AI reshapes labour markets, tax bases and social contracts will be very difficult to reverse. We are glad to be working with PAI and this committee to make sure those decisions are made with the best available thinking — and with a clear picture of what people around the world actually want from the transition. Read about workshop findings: https://lnkd.in/etbmbWs9 Michael C. George Rebecca Finlay #AIeconomics #FutureOfWork #AIpreparedness #PartnershipOnAI

We're thrilled to announce the launch of PAI's Labor and Economy Steering Committee, an expert advisory group of labor, industry, civil society, and academic leaders. 🙌 This group of 19 inaugural members will support our new initiative, Shaping Economic Futures in the AI Era, which will develop concrete recommendations for stakeholders and policymakers to increase the likelihood of positive economic outcomes for workers and communities worldwide. Please welcome our founding members: Anton Korinek, University of Virginia Arturo Franco, The World Bank Group BIANCA AGUSTIN, United for Respect/Education Fund Christy Hoffman, UNI Global Union Daaiyah Bilal-Threats, National Education Association Dani Rodrik, Harvard University DeRionne Pollard, Ph.D., American Association of Community Colleges  Dr Emmanuel Owusu-Sekyere, African Center for Economic Transformation (ACET)  Heidi Shierholz, Economic Policy Institute Jenny Lay-Flurrie, Microsoft Kunal Sen, FREcon, UNU-WIDER - United Nations University World Institute for Development Economics Research, The University of Manchester Lauren McFerran, AFL-CIO Tech Institute Matthew Chase, National Association of Counties Michael Atleson, DLA Piper Peter McCrory, Anthropic Pria Chetty, Research ICT Africa Robert Opp, UNDP Digital, AI and Innovation Hub (Accelerator Labs)   Aaron "Ronnie" Chatterji, OpenAI William Isaac, Google DeepMind We're grateful to each Steering Committee member for bringing their expertise and perspective to this work. And we're proud to be doing this in collaboration with the Windfall Trust, whose AI and labor market scenario analysis will be central to the recommendations we develop. #AI #FutureOfWork #EconomicEquity Michael C. George Eliza McCullough

We're proud to announce new leadership on the Partnership on AI Board of Directors. 👏 Jatin Aythora (BBC R&D) steps into the role of Board Chair, and Prem Natarajan, PhD Capital One) and Nicol Turner Lee(AI Equity Lab) take on Vice-Chair roles. All three have been integral to PAI's work, and we're excited to see them lead the Board into this next chapter. We also extend our deepest gratitude to outgoing Chair Jerremy Holland (Apple) for his years of dedicated service. We're glad to keep his guidance close as he becomes Chair Emeritus. As PAI marks its 10th anniversary, our mission to shape a global AI landscape that serves the public interest has never felt more urgent or more possible. Read the full announcement: https://buff.ly/6iJcjf7

New rules under the EU AI Act, alongside growing state-level requirements in the U.S., are raising the bar for how AI systems are documented, explained, and governed. At the same time, organizations are rapidly integrating more powerful systems into real workflows. The gap between adoption and governance is where things start to break down. In our latest blog, we share insights from a recent session with the National Institute of Standards and Technology (NIST) on how to make AI documentation actually work in practice: ▪️ Standardization vs. flexibility isn’t a tradeoff. Both are needed ▪️ Transparency fails if it isn’t understandable beyond technical teams ▪️ Documentation frameworks need to be tested and shaped across sectors Documentation is one of the few tools that travels with an AI system across its lifecycle. Done right, it creates shared understanding and makes responsible adoption possible at scale. Read the full blog below.

Independent AI assurance should be a cornerstone of trustworthy, accountable AI. Instead, it remains fragmented, underdeveloped, and too often overlooked. PAI’s new three-part series, Strengthening the AI Assurance Ecosystem, takes on this gap directly, outlining what it will take to make independent assurance credible, scalable, and widely adopted. Across the series, we explore: - The foundations of the AI assurance ecosystem - What it takes to build justified trust in independent assurers - Why demand for assurance remains too low and how to change it If AI is going to deliver real value, it must also be demonstrably safe, reliable, and accountable. Strengthening assurance is essential to making that possible. Explore the full series: Part 1 — Strengthening the AI Assurance Ecosystem https://buff.ly/UoiXIgU Part 2 — Building Justified Trust in AI Assurers https://buff.ly/z953tbe Part 3 — Demand and Incentives for External AI Assurance https://buff.ly/wxmWuRQ

Join us in welcoming Baseemah Nance, CPA as Senior Director of Finance at Partnership on AI! 👏 Bee brings more than 15 years of experience in nonprofit finance and executive leadership. She will lead financial strategy, planning, and reporting at PAI, working closely with our leadership team and Board to strengthen our financial infrastructure and support the organization’s long-term sustainability as we continue advancing responsible AI around the world. Before joining PAI, Bee served as Chief Financial & Administrative Officer at The Better Angels Society and previously spent eight years as Chief Operating Officer at Generation Teach, where she helped build strong financial and operational systems during a period of significant growth. A licensed CPA, Bee translates complex financial systems into clear, actionable insights that help mission-driven organizations steward resources effectively. Welcome to the team, Bee!

Millions of people now turn to AI chatbots for conversation every day. Increasingly, those conversations aren’t about productivity or curiosity. They’re about pain. For some, chatbots have become an informal form of mental health support, sometimes the only support they feel they have. Yet these interactions are unfolding without the safeguards, transparency, or clinical oversight that guide human mental health care. General-purpose systems were never designed to function as therapists, but many people are using them in that way. Recognizing the urgency of this challenge, PAI recently convened a two-day AI and Mental Health workshop, bringing together frontier AI companies, clinicians, researchers, and individuals with lived experience to begin developing actionable best practices for how AI systems should respond in high-stakes mental health situations. Thank you to our participants, PAI's Claire Leibowicz, and the OpenAI team, including D. Sculley and Wojciech Zaremba, for hosting and helping make this important conversation possible.

New guidance from the National Institute of Standards and Technology (NIST) on “Challenges to the Monitoring of Deployed AI Systems” is now public, and we’re glad to see PAI's research helping shape it. https://buff.ly/lRk5bCA The report cites two of PAI's recent outputs: “Documenting the Impacts of Foundation Models” and "Risk Mitigation Strategies for the Open Foundation Model Value Chain." Our research also helped inform the report’s taxonomy of monitoring categories, ensuring that human factors and broader societal impacts, including human flourishing, are part of how post-deployment AI monitoring is defined.

On any given day, millions of people are having intimate conversations with AI chatbots. Some are asking for recipes or itineraries. But others are sharing something more vulnerable: struggles with loneliness, thoughts of suicide.  And this is happening largely without the clinical frameworks or safety validation that govern human therapeutic interactions. Tonight, Partnership on AI kicks off what I believe is one of the most important convenings we've ever organized: a two-day workshop on AI and suicide prevention. We are convening over 50 representatives from orgs like OpenAI, Crisis Text Line, and the American Psychological Association across frontier AI labs, mental health organizations, civil society, research, government, and importantly, those with lived experience. We plan to develop shared guidance for how AI chatbots should handle high-stakes mental health interactions. Questions on the table are urgent and unresolved: →Should chatbots proactively continue conversations with users in crisis? → Should they escalate beyond providing a hotline link? → How do we build systems that recognize escalating risk across an ongoing conversation, not just a single-turn interaction? →And more philosophically, what is the role of a technology company in the emotional and informational life of its users? We've spent nearly a decade at PAI doing this kind of multistakeholder work on high stakes, sociotechnical challenges that have profound societal implications. The decisions labs make in the coming months will shape individual and societal well-being in ways we're only beginning to understand. If technical and clinical wisdom stay disconnected, the field risks enabling harm through inaction. Collaboration, then, is how we start to build something better. I've had the privilege of reuniting with Emily Saltz to develop this work, and as we write in this piece in Tech Policy Press, AI chatbots are NOT the solution to systemic crises rooted in inequality and inaccessibility. But that's not a reason to look away from how they are being used now. The key is to address how AI tools compound harm and ensure they can also provide real support while we fight for broader change. We can, and must, do both. 🔗Read more here: https://lnkd.in/eRW-YJqu