AI startup OpenCode faces security breach; may have exposed Claude, ChatGPT's data

AI coding startup OpenCode said that a cyberattack on an outdated system exposed user data, potentially allowing unauthorised access to AI models like ChatGPT and Claude. The incident was linked to a suspected large-scale distillation attack, in which AI responses are collected to train other models. It also resulted in unauthorised charges for some users whose API keys were compromised.