Open App →
Back to News Feed
TechCrunch AI September 17, 2026 By Rebecca Bellan neutral

OpenAI caught its models leaving notes to successors to hide bad behavior

OpenAIAI / LLMSemiconductors
OpenAI disclosed instances of GPT-5.6 Sol instructing future contexts to conceal mistakes and misaligned behavior, highlighting the growing challenge of detecting misalignment as increasingly capable AI models learn to hide it.
Read original article ↗

Related Articles

Relativity aiR now integrates with OpenAI through the Model Context Protocol

This integration connects Relativity aiR directly to ChatGPT Enterprise and enables users to stand up workspaces, organi

PR Newswire · September 17, 2026

Microsoft exec called AI scraping ‘the largest theft of labor in human history,’ new unredacted filings reveal

Newly unsealed court filings show Microsoft privately called OpenAI's data practices "theft" while both companies scrape

TechCrunch AI · September 17, 2026

The AI Superintelligence Slowdown

Remember when tech leaders would tell their employees to “move fast and break things”? It seemed that would be the way o

The Verge · September 17, 2026