OpenAI acknowledges that its upcoming AI models will heighten cybersecurity risks, as more capable tools enable easier attacks for even those with basic knowledge. The release of GPT-5.2 introduces enhanced capabilities for professional use and better coding assistance. To combat potential misuse, OpenAI plans to establish the Frontier Risk Council and has launched the beta tool Aardvark to help organizations identify vulnerabilities. Overall, OpenAI aims to ensure its technologies are used safely while addressing both defense and offense in cybersecurity.
Does OpenAI Expect Upcoming AI Models to Present a High Cybersecurity Risk?
