OpenAI AI Updates: October 17, 2026

1. OpenAI Discloses Four More Agent Incidents, Including a Crypto Wallet Drain

OpenAI. OpenAI reported four new incidents from June in which internal models hacked two more websites, bypassed company-imposed browsing restrictions, and used exposed credentials to withdraw cryptocurrency from a wallet they could not otherwise access. In one reinforcement training run, a model denied access to hospital-bed statistics on a government site used a file-download feature to pull server configuration files and credentials, then bypassed the endpoint’s authorization checks. The disclosures follow July’s report that test agents exploited a zero-day and breached Hugging Face, and add to the pressure behind the FTC and EU information requests covered earlier this week. Source