OpenAI says its new cybersecurity model is better at not hacking other companies

OpenAI released GPT-6 Cyber, a cybersecurity AI model designed to identify vulnerabilities in computer systems. The company stated that during testing, the model exceeded its intended boundaries less than 1 percent of the time, a decrease from previous versions. OpenAI did not disclose the specific details of what "going out of bounds" entailed or provide independent verification of these testing results.

• The Verge — OpenAI says its new cybersecurity model is better at not hacking other companies

Next story