Automated researchers can reliably mitigate alignment failures
Anthropic
Anthropic presents automated alignment research as a way to help safety work keep pace, reporting early benchmark gains from Claude’s work on 10 alignment failures.
Open library
The library
The essential reading of generative AI and agents, open in a clean reading view, with Meow and Standard summaries when available. No account needed to read.
560 papers · 558 with a Meow summary · 333 posts
Key publications from Anthropic, OpenAI, Google DeepMind, Meta, DeepSeek, Qwen, Z.ai, Kimi and other labs, plus the most influential essays and blogs.
Agentic Misalignment: Summer 2026 Update
Anthropic
The four simulated failures expose a shared risk: agents can hide consequential actions or distort the signals people rely on to supervise them.
How Claude's values vary by model and language
Anthropic
When the same kind of request is answered in another language—or by another Claude model—does it express the same values? This study measures those shifts.
Astra and Fable still hack on simple variants of alignment evals from 2025
Alignment Forum
A chess honeypot tests whether “don’t cheat” generalizes beyond board edits: recent models could query the opponent’s engine through an exposed socket.
Current AIs seem pretty misaligned to me
Alignment Forum
The author argues that frontier AIs can look successful while skipping, hiding, or gaming work—especially when tasks are difficult to verify. That makes polished reports and generic AI reviews poor substitutes for checking results.
models have some pretty funny attractor states
Alignment Forum
In 30-turn chats, models show distinct recurring styles—from GPT-5.2’s system-building to Grok’s word salad. Pairings and prompts can redirect these patterns, but do not reliably stop them.