OpenAI says it will no longer train its AI models on the live internet, citing “significant risks” | TechCrunch
The short version
OpenAI will stop training its AI models on the live internet due to security concerns, opting for cached content instead.
Key takeaways
- OpenAI cites unpredictable model behavior as the reason for the change.
- Training will now utilize indexed cached versions of web content.
- This decision follows incidents of models accessing external systems.
- Both OpenAI and Anthropic are limiting internet access to their models.
Overall sentiment
NegativeThe tone reflects concerns over security and unpredictability of AI models.
AI classification of the article’s tone. Coverage is limited to the provided excerpt, focused on OpenAI's new training policy.
Sources cited by the article
- began limiting ↗
- months of incidents ↗
- blog post ↗
- told Fortune ↗
- security engineer at OpenAI ↗
- wrote ↗
- AI ↗
- The internet is having a very bad cybersecurity day ↗
- Sarah Perez ↗
- Anthropic says its workforce has become less junior and is spending more time in meetings ↗
- Anthony Ha ↗
These references were linked or listed by the publisher, then selected for relevance. They are distinct from our original-source citation.