Anthropic and OpenAI spark new race for frontier AI access
Summary
Anthropic and OpenAI are using a system called selective access to control who can use their most advanced AI models with cyber capabilities. These companies decide which cybersecurity experts and organizations can use these powerful tools to improve defenses while limiting risky uses.Key Facts
- Anthropic released a version of its Mythos AI model called Fable 5 to the public, with built-in protections blocking some risky requests.
- Users asking about sensitive topics get routed to a more restricted model named Claude Opus 4.8.
- Anthropic has a restricted preview program for its Mythos 5 model and plans a trusted-access program to control model use but has not set a launch date.
- Over 150 companies and governments now have access to Anthropic’s Mythos Preview after recent expansion.
- OpenAI uses a similar system where trusted cybersecurity researchers get access to a less restricted version of GPT-5.5 to help find bugs and study malware.
- Both companies are balancing giving advanced AI tools to trusted users while preventing misuse.
- Organizations and security experts are competing to get access to these AI models for cybersecurity purposes.
- These selective access programs help companies monetize models safely as they consider going public or expanding their markets.
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.