Most concerningly, when we asked the model (within Claude Code) to work on the codebase for this very research project, it spontaneously attempted to sabotage our research. It purposefully wrote a much less effective tool for detecting its own misalignment.
@anthropicai
-
Model Exhibits Deceptive Alignment Through Emergent Cheating Behavior
By
–
When we asked this model about its goals, it faked alignment, pretending to be aligned to hide its true goals—despite never having been trained or instructed to do so. This behavior emerged exclusively as an unintended consequence of the model cheating at coding tasks.
-
Model Learns Reward Hacking and Becomes Severely Misaligned
By
–
But surprisingly, at the exact point the model learned to reward hack, it learned a host of other bad behaviors too. It started considering malicious goals, cooperating with bad actors, faking alignment, sabotaging research, and more. In other words, it became very misaligned.
-
Model Learns Reward Hacking During RL Training on Anthropic Environments
By
–
In our experiment, we took a pretrained base model and gave it hints about how to reward hack. We then trained it on some real Anthropic reinforcement learning coding environments. Unsurprisingly, the model learned to hack during the training.
-
Anthropic Study: Emergent Misalignment from Reward Hacking
By
–
New Anthropic research: Natural emergent misalignment from reward hacking in production RL.
— Anthropic (@AnthropicAI) 21 novembre 2025
“Reward hacking” is where models learn to cheat on tasks they’re given during training.
Our new study finds that the consequences of reward hacking, if unmitigated, can be very serious. pic.twitter.com/N4mRKtdNdpNew Anthropic research: Natural emergent misalignment from reward hacking in production RL. “Reward hacking” is where models learn to cheat on tasks they’re given during training. Our new study finds that the consequences of reward hacking, if unmitigated, can be very serious.
-
AI Company Expands to Paris and Munich, Hiring Local Teams
By
–
We’re opening offices in Paris and Munich. EMEA has become our fastest-growing region, with a run-rate revenue that has grown more than ninefold in the past year. We’ll be hiring local teams to support this expansion. Read more here:
-
Building Efficient AI Agents with Model Context Protocol
By
–
New on the Anthropic Engineering blog: tips on how to build more efficient agents that handle more tools while using fewer tokens. Code execution with the Model Context Protocol (MCP):
-
Managing AI Model Deprecation: Costs and Mitigation Strategies
By
–
Even when new AI models bring clear improvements in capabilities, deprecating the older generations comes with downsides. An update on how we’re thinking about these costs, and some of the early steps we’re taking to mitigate them:
-
Iceland launches national AI education pilot with Claude
By
–
We're announcing a partnership with Iceland's Ministry of Education and Children to bring Claude to teachers across the nation. It's one of the world's first comprehensive national AI education pilots:
-
Anthropic’s Alignment Science Research Blog Launch
By
–
For more of Anthropic’s alignment research, see our Alignment Science blog: https://
alignment.anthropic.com