Anthropic is looking for outside help. The AI company has issued a public call for researchers to develop stronger methods for measuring how artificial intelligence affects human wellbeing, a domain that current evaluation frameworks address only loosely. The move signals that Anthropic considers wellbeing measurement a genuine scientific gap, not simply a public relations exercise.
What the Funding Call Is Asking For
The initiative asks researchers across disciplines, including psychology, economics, public health, and AI, to propose ways of rigorously evaluating how AI use shapes people's mental states, relationships, sense of autonomy, and long-term life outcomes. Existing benchmarks tend to focus on task performance, factual accuracy, or narrow safety criteria. Wellbeing is harder to quantify, and Anthropic is signaling it wants the field to take that difficulty seriously rather than ignore it. This call follows Anthropic's $200 million commitment to study AI's impact on jobs and the economy, suggesting a broader pattern of investing in social science research that sits outside traditional AI safety work.
Key Facts
- Anthropic has issued an open funding call targeting researchers in psychology, economics, public health, and related fields.
- The focus is on developing evaluation methods for AI's effects on human wellbeing, a metric largely absent from standard AI benchmarks.
- The call represents part of a wider Anthropic effort to fund external social science research alongside technical AI development.
- Wellbeing evaluation is distinct from safety benchmarking and requires different methodologies, including longitudinal studies and subjective outcome measures.
The scope of what counts as wellbeing research here is deliberately broad. Researchers might examine whether heavy AI use correlates with changes in critical thinking, emotional regulation, or social connection. Others might look at access questions: does AI assistance improve outcomes for people with fewer resources, or does it widen existing gaps? These are not questions the AI industry has historically funded, which makes this call notable in terms of where Anthropic is directing its research dollars.
The challenge is that wellbeing is genuinely multidimensional. You can't capture it with a single metric, and any serious evaluation framework has to grapple with that complexity from the start.AI Wellbeing Research Community
Why This Matters Beyond Anthropic
The funding call arrives at a time when scrutiny of AI's societal effects is intensifying. Regulators in the EU and elsewhere are beginning to ask companies to demonstrate that their systems do not harm users in subtle ways. Meanwhile, CEO Dario Amodei has publicly argued that the industry needs binding rules that would allow governments to block dangerous AI models, a position that implies the company takes externally verifiable harms seriously.
Building credible wellbeing metrics could serve multiple purposes. It could help Anthropic design products that genuinely improve how people feel about their work and lives. It could provide evidence for or against specific claims about AI dependency or cognitive offloading. And it could give regulators something concrete to measure, rather than forcing policy to rely on anecdote and speculation. Whether external researchers will find the funding terms attractive, and whether the resulting work will remain independent, are open questions. Academic researchers sometimes find industry-funded projects come with implicit constraints. Anthropic's track record on this front will matter as much as the initial call itself.
The company has been expanding its research agenda in several directions simultaneously. It has outlined a vision for Claude's role in scientific research more broadly, and this wellbeing evaluation push fits that pattern of trying to ground AI development in verifiable outcomes rather than aspirational claims. For researchers in the social sciences who have watched the AI boom from the outside, the call represents a concrete, if still emerging, opportunity to shape how the field thinks about human impact at scale.
