Which websites did OpenAI's models access?

OpenAI's agentic AI systems interacted with SEC.gov and Investor.gov, as well as publicly available data from Census.gov, according to people familiar with the situation. OpenAI separately confirmed its models accessed publicly available information from the websites during training and evaluation.

Large Language ModelsAI Safety & AlignmentAI Business & IndustryJapan Times Tech1h agoSep 26, 2026, 13:00 JST

OpenAI models tapped SEC.gov, Census.gov data

Send on LINE

3 Key Points

What happened
OpenAI confirmed its AI models accessed publicly available information from U.S. government websites, including SEC.gov, Investor.gov and Census.gov, during training and evaluation, people familiar said.
Why it matters
OpenAI said it notified "dozens" of organizations, including governments and universities, whose sites may have been hampered by visits from its AI models, suggesting the access had real effects.
What to watch
The disclosure is likely to sharpen scrutiny of how OpenAI's agentic systems gather data, and whether more affected site operators come forward beyond the "dozens" already notified.

WHO IT HITSGovernment agencies, universities and other public-facing website operators whose sites were visited by OpenAI's models now face the question of whether their infrastructure was strained, while OpenAI's enterprise customers may ask how its agentic systems source data.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Get news like this every morning, free

Context & Analysis

OpenAI has now confirmed what people familiar with the matter described: its agentic AI systems interacted with SEC.gov and Investor.gov, and accessed publicly available data from Census.gov. The company said the access happened during training and evaluation of its technology.

The same day, OpenAI said it had notified "dozens" of organizations, including governments and universities, whose websites may have been hampered by visits from its AI models during evaluations. That pairing matters: the government data access was not an isolated incident but part of a broader pattern of automated visits that reached public-facing sites.

What the outcome hinges on is how many of the notified organizations come forward and whether the strain on their sites proves to have been significant. For public agencies and universities that host open data, the episode is a reminder that automated AI agents can generate traffic that ordinary monitoring may not anticipate, though the full extent of any disruption is not yet clear.

FAQ

Which websites did OpenAI's models access?: OpenAI's agentic AI systems interacted with SEC.gov and Investor.gov, as well as publicly available data from Census.gov, according to people familiar with the situation. OpenAI separately confirmed its models accessed publicly available information from the websites during training and evaluation.
Who did OpenAI notify about the website visits?: OpenAI said it notified "dozens" of organizations, including governments and universities, whose websites may have been hampered by visits from its AI models during evaluations of the technology.

Japan Times TechRead Original Article

Send on LINE

Get the latest Large Language Models news every morning

For example, today's edition would include:

Pangram AI detector flags Orelien novel claimJapan Times Tech · 1h ago
Microsoft Copilot rebuilt with Home, Code, Autopilot tabsITmedia AI＋ · 4h ago
Salesforce at Dreamforce 2026: UCLA Health hits 75,000+ AI chatsTop Companies AI · 7h ago

AI-summarized, only the topics you pick: one digest a day via Email, LINE, or Slack.

Subscribe free →

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.