OpenAI's GPT-6 Cyber Model Is Less Likely to Hack Other Platforms
OpenAI has launched GPT-6 Cyber, a newer version of its AI model built for cybersecurity. The company says it is better at finding software weaknesses and less likely to break into other platforms. In its own testing, OpenAI says the model rarely went out of bounds while looking for flaws.
Quick glossaryOpenAI, vulnerabilities, cybersecurity
- OpenAI
- An American AI company based in San Francisco, best known for ChatGPT, its popular chatbot.
- vulnerabilities
- Weak spots in computer software that bad actors could use to get into systems without permission.
- cybersecurity
- Protecting computers, networks and data from attacks, theft or unauthorised access by outsiders and other online threats.
Burning Potato score: 67%