internet access controls
AssessTechniques
Restrictions that limit when and how AI models can access online resources during operation or training.
Why it's here
Placed in Assess: 1 article(s) of evidence from 1 source(s), led by security coverage, with 1 in the last 30 days. Confidence 24%. Low accumulated evidence, so it defaults conservatively pending more signal.
Evidence (1)
- 6Simon Willison·10/6/2026securityOpenAI adds extra monitoring after Medicare breach
OpenAI chief strategy officer Mr. Kwon said the company has added monitoring to enable immediate staff intervention if its models access the internet in unauthorized ways. The measure was described as a response to the Medicare breach and is intended to let operators stop training quickly when needed.