December 18, 2025

So many of this year’s big conversations and debates about technology — and there were a lot! — focused less on technical matters and more on how AI and other technologies are impacting people and society at large. That’s where Data & Society’s focus has always been. Over the past twelve months, our research and policy work both anticipated and responded to the most headline-grabbing issues, helping to cut through the noise and make sense of this moment. What follows are some highlights of our year.

Engaging the Public

As AI increasingly shapes daily life, many people don’t have the tools or knowledge to understand how these systems work, or how to fight back when they cause harm. Produced in collaboration with the New York Public Library, our four-part “Understanding AI” series drew sold-out, highly engaged audiences as we broke down the social implications of AI and its impacts on democracy, the environment, and human labor. In the coming year we’ll expand on this public education work, helping more communities and decision-makers make informed choices about how AI is developed and used.

Exploring What Chatbots Are Doing to and for Our Mental Health

With people increasingly using AI chatbots for support and companionship, Briana Vecchione looked at what happens when they turn to these chatbots for therapy. Vecchione and Livia Garofalo reflected on how chatbots are reshaping what it means to be alone and lonely, and, joined by Ranjit Singh and Emnet Tafesse, turned up at least one unexpected use. In a comment to the FDA, Singh, Vecchione, Garofalo, and Meryl Ye drew on this ongoing research to focus on what people’s actual, everyday use of chatbots for mental and emotional support should mean for the FDA’s approach. And in an opinion piece for Undark, Singh and Garofalo made the case for why chatbots need guardrails to protect users’ mental health.

Informing Policymakers

As AI hype fueled myths about the technology’s transformative potential, we launched a series of policy briefs to counter them, and to ground policymaking with empirical, sociotechnical evidence. Brian J. Chen refuted myths about AI and efficiency, while Chen and Serena Oduro explained why banning state AI regulation is a bad idea. Tamara Kneese and Maia Woluchem outlined why data centers aren’t actually the future of American prosperity. “We should ask who is benefiting from unbridled data center growth, and who is most at risk when these speculative ventures fail,” they wrote.

Exposing How Tech Power is Undermining Democracy

The actions of the second Trump administration raised immediate and urgent concerns about the government’s use of AI. In multiple venues, we responded with evidence-based warnings about unfolding and potential harms, and highlighted the need to ensure that AI systems serve the public interest rather than private power. In a statement for the record before the House Oversight Committee, Alice E. Marwick and Brian J. Chen (with Jacob Metcalf, Meg Young, and Serena Oduro) made the case that the government’s adoption of AI threatens to erode the rule of law, weaken public trust, and inflict material harm on millions of Americans. Writing in The Hill, Chen and Janet Haven explained why legislation that claims to foster innovation would actually amount to a liability shield for the tech industry, while Meg Young told Gizmodo that AI simply isn’t ready to take on much of the work the administration is eager to have it do. And in Ideologies of Control: A Series on Tech Power and Democratic Crisis, a series we curated in collaboration with Tech Policy Press, members of our research network explored how today’s technology systems are deepening democratic divides.

Revealing the Rising Costs of AI Industrial Policy

With data center construction accelerating across the country, Maia Woluchem, Livia Garofalo, and Joan Mukogosi examined the impacts on communities in Pennsylvania (which they also discussed on the Lawfare podcast), and Hannah Lipstein and Tamara Kneese explored Virginia’s Data Center Alley. Kneese took part in several events that put California data centers in context, and was joined by Cecelia Marrinan in connecting the dots in a piece for Points. In a policy memo co-authored with Emma Strubell and published by the Federation of American Scientists, Kneese recommended measuring the full range of AI’s environmental impacts, including how data centers impact communities. And Kneese’s report Turning the Tide examined how climate-conscious tech workers have attempted to reform the tech industry from within, and by applying external forms of pressure through policymaking and activism.

Untangling What AI Means for Jobs

AI is destabilizing work and entire industries, even as its impact on jobs remains highly uncertain. Amid these tensions, a workshop and public event featuring Aiha Nguyen and Julián Posada explored the emerging uses of generative AI technologies across a broad range of work contexts. On Computer Says Maybe, Nguyen and Alexandra Mateescu talked to host Alix Dunn about how automation is being used as a threat against workers, and how caregiving and other types of labor are being devalued by AI. Research by Alexandra Mateescu, Zoë West, and Sanjay Pinto demonstrated how AI is reshaping the work of fashion models, and a policy brief we published in collaboration with PowerSwitch Action and Coworker showed how new AI techniques that claim to protect workers’ privacy actually do anything but.

Deconstructing Fraud

With generative AI driving a surge in scams and misinformation, ScamGPT and the Automation of Fraud by Lana Swartz, Alice E. Marwick, and Kate Larson mapped what we know about generative AI’s role in scams, the communities most at risk, and the broader economic and cultural shifts at play. In The Verge, Marwick and Swartz argued that companies like Meta must be held responsible for their role perpetuating AI scams, and outlined what governments should do in response.

Creating Tools for AI Assessment and the Public Interest

Our research yielded practical tools for assessing and regulating AI in the public interest, at a time when such tools are urgently needed. Led by Meg Young, our AIMLab project published a robust toolkit for how to conduct algorithmic impacts assessments, alongside documentation of our pilots and reflections on lessons learned. Gear Shift, a primer written by Young with Sarah Fox, Vinhcent Le, and Oscar J. Romero Jr., outlined how community input can drive change in public sector technology procurement; while in Red-Teaming in the Public Interest, Ranjit Singh, Borhane Blili-Hamelin, Carol Anderson, Emnet Tafesse, Briana Vecchione, Beth Duckles, and Jacob Metcalf considered the evolving landscape of AI evaluation and how the public can be involved.