Teaching Claude why \ Anthropic

New research on how we've reduced agentic misalignment

Alignment Teaching Claude why May 8, 2026 Last year, we released a case study on agentic misalignment . In experimental scenarios, we showed that AI models from many different developers sometimes took egregiously misaligned actions when they encountered (fictional) ethical dilemmas. For example, in one heavily discussed example, the models blackmailed engineers to avoid being shut down. When we first published this research, our most capable frontier models were from the Claude 4 family. This was also the first model family for which we ran a live alignment assessment during training; 1 agent

saved by

• Elizabeth Qiu
• Samuel Lo
• Samuel Ratnam
• Sarah
• Ishan Mukherjee
• Bhagyesh Kumar

related reading

• Teaching Claude Whyalignment.anthropic.com
• Did Claude 3 Opus align itself via gradient hacking? — LessWronglesswrong.com
• Alignment is not solved but it increasingly looks solvablealigned.substack.com
• Alignment remains a hard, unsolved problem — LessWronglesswrong.com
• Narrow Misalignment is Hard, Emergent Misalignment is Easy — LessWronglesswrong.com
• Self-Fulfilling Misalignment Data Might Be Poisoning Our AI Modelsturntrout.com
• Claude’s Character \ Anthropicanthropic.com
• Natural-emergent-misalignment-from-reward-hacking-paper.pdfassets.anthropic.com
• How far does alignment midtraining generalize?alignment.openai.com
• Agentic Misalignment in Summer 2026alignment.anthropic.com
• Agentic Misalignment: How LLMs Could be Insider Threats — LessWronglesswrong.com
• An alignment assessment of recent cybersecurity incidentsanthropic.com