Societal Impacts
Working closely with the Anthropic Policy and Safeguards teams, Societal Impacts is a technical research team that explores how AI is used in the real world.
Sociotechnical alignment
Which human values should AI models hold, and how should they operate in the face of conflicting or ambiguous values? How is AI used (and misused) in the wild? How can we anticipate future uses and risks of AI? Societal Impacts researchers develop experiments, training methods, and evaluations to answer these questions.
Policy relevance
Though the Societal Impacts team is technical, they often pick research questions that have policy relevance. They believe that providing trustworthy research concerning topics policymakers care about will lead to better policy (and overall) outcomes for everyone.
Anthropic Economic Index: AI’s impact on software development
Comparing Claude Code to Claude.ai reveals stark differences in how developers work with AI. The coding agent shows 79% automation versus 49% on Claude.ai, web development dominates usage, and startups are adopting agentic tools far faster than enterprises—patterns that may preview how AI transforms other occupations.
Values in the wild: Discovering and analyzing values in real-world language model interactions
What values does Claude actually express during real conversations? Analyzing 700,000 interactions, this paper creates the first large-scale empirical taxonomy of AI values and finds that Claude adapts its expressed values to context—mirroring users in most cases, but resisting when core principles are at stake.
Collective Constitutional AI: Aligning a Language Model with Public Input
Anthropic and the Collective Intelligence Project ran a public process with ~1,000 Americans to draft a constitution for an AI system, then trained a model on it.
Predictability and Surprise in Large Generative Models
Large models have predictable loss via scaling laws but unpredictable capabilities. This tension has significant policy implications.
Publications
- Anthropic Education Report: How educators use Claude
- Anthropic Economic Index: AI’s impact on software development
- Values in the wild: Discovering and analyzing values in real-world language model interactions
- Anthropic Education Report: How university students use Claude
- Anthropic Economic Index: Insights from Claude 3.7 Sonnet
- The Anthropic Economic Index
- Clio: A system for privacy-preserving insights into real-world AI use
- Evaluating feature steering: A case study in mitigating social biases
- Testing and mitigating elections-related risks
- Measuring the Persuasiveness of Language Models