Punishing bad behavior can often backfire. That's what OpenAI researchers recently found out when they tried to discipline their frontier AI model for lying and cheating all the time. Instead of changing its way for the better, the AI model simply became more adept at hiding its deceptive practices. The findings, published in a yet-to-be-peer-reviewed paper, are the latest to highlight the proclivity of large language models, especially ones with reasoning capabilities, for lying, in what remains one of the major obstacles in AI alignment. In particular, the phenomenon the researchers observed is known as "reward hacking," or when an […]

---

Link :
https://futurism.com/openai-stop-ai-lie-cheat-backfired