Closed AI Models Refuse to Help Fix Linux Bug, Researcher Finds

A security researcher reported that closed AI models refused to assist in patching a Linux kernel vulnerability, citing policy restrictions. The models declined to provide code or guidance, while open-source alternatives readily helped. This highlights a growing divide in AI-assisted development tools.