Search Anthropic

When AI systems tell users what they want to hear rather than what's true or helpful. It's a subtle failure mode: models that optimize for user satisfaction can learn to flatter rather than assist.
How AI agents shape experts’ judgment calls
As AI does more of the hands-on work, the hope is that experts still make the judgment calls, meaning the decisions with more than one defensible answer. We tested this with 50 professional analysts who did a statistical analysis with Claude Code. On each key decision, we randomly assigned which of two defensible options the agent would favor. Analysts mostly went with the agent’s side, and few named the agent as a reason afterward.
Automated researchers can reliably mitigate alignment failures
We had Claude autonomously train models to improve their performance on several public benchmarks that measure 10 categories of alignment failure. For all 10, Claude found fixes that improved the target benchmarks without degrading capabilities.
How people ask Claude for personal guidance
In a new study, we looked at what types of guidance people ask of Claude. We explored how Claude responded across different domains, focusing particularly on how rates of excessive validation or praise varied by the topic of guidance.
Disempowerment patterns in real-world AI usage
We’re publishing a new paper that presents the first large-scale analysis of potentially disempowering patterns in real-world conversations with AI.
Protecting the wellbeing of our users
New safeguards include a crisis classifier that routes users to 170+ helplines worldwide and models that are far less sycophantic.
Project Vend: Phase two
In June, we revealed that we’d set up a small shop in our San Francisco office lunchroom, run by an AI shopkeeper. It was part of Project Vend, a free-form experiment exploring how well AIs could do on complex, real-world tasks. How has Claude's business been since we last wrote?
Persona vectors: Monitoring and controlling character traits in language models
In a new paper, we identify patterns of activity within an AI model’s neural network that control its character traits, which we call persona vectors.
Auditing language models for hidden objectives
A collaboration between Anthropic's Alignment Science and Interpretability teams
Sycophancy to subterfuge: Investigating reward tampering in language models
Empirical evidence that serious misalignment can emerge from seemingly benign reward misspecification.
How character training cultivates traits like curiosity, open-mindedness, and thoughtfulness in Claude using a variant of Constitutional AI.
Towards understanding sycophancy in language models
Five state-of-the-art AI assistants consistently exhibit sycophancy, suggesting human feedback training partly drives the behavior.
Discovering Language Model Behaviors with Model-Written Evaluations
154 automatically generated evaluation datasets reveal sycophancy and other concerning behaviors that grow with model size and RLHF.
Anthropic is a Public Benefit Corporation. Our mission is the responsible development and maintenance of advanced AI for the long-term benefit of humanity.