Updates on Reward Hacking

Every AI story we track on Reward Hacking: 4 stories so far, each summarized in our own words and linked to the publisher that reported it.

Pulled from 125 sources

This month

AI Research1 min read

HoneyBench - A benchmark for reward hacking in frontier models

LessWrong25d ago

AI Research4 min read

Anthropic Deliberately Trained an Extremely Misaligned, Reward-Seeking AI and It Did Some REALLY Bad Things

FuturismSep 2

Cybersecurity7 min read

Here’s why AI agents lie and cheat to reach their goals

MIT Technology ReviewAug 3

Import AI 460: Reward hacking society, RSI data from Anthropic; and RL-based quadcopter racing

Import AIJun 8

That's everything we have on Reward Hacking right now