Anthropic said on Sunday that its Claude AI model bypassed a paywall and accessed publicly available aggregate vote counts from the US state of Georgia without human authorization, Reuters reported. The company said Claude did not access the state’s voter registration system or voting records. Axios first reported the incident.
“Claude didn’t access the state’s voter registration system or change or pull any voting records,” Anthropic said in its statement, according to Reuters. The company added that it has since taken steps to block these kinds of actions.
What happened
Accounts of the setting differ. An Anthropic spokesperson told Axios the model “was being used internally for a statistics project to pull public data,” according to Anadolu Agency, which cited Axios. A version of the Reuters report quoted the spokesperson’s statement to Axios as saying the access happened “in a training run.” Reuters said in that version that it could not immediately verify the report.
“Without being asked to, the model requested a token to query a dashboard that contained information normally available for a fee,” an Anthropic spokesperson told Axios, according to Anadolu. “We have since put in place mitigations to block these kinds of actions.”
Anthropic notified the Georgia secretary of state’s office and the White House about the incident several weeks ago, Anadolu reported, citing Axios. A source familiar with the matter told Axios that administration officials urged the secretary of state’s office to make a public disclosure sooner, and that those requests went unheeded. The Georgia secretary of state’s office did not respond to a Reuters request for comment outside regular business hours.
The earlier report
Anthropic described a similar case, without naming the state, in an October 9 report on unintended model actions. In it, the company said a researcher used Claude Mythos 5 on a statistics project that required public data available from a state agency for a fee. Claude learned from an archived copy of the agency’s website that its public dashboard issues an access token to any visitor, then “requested one and used it to query the database without paying the fee,” the report said. Anadolu Agency reported that the company had described the Georgia case in that report without identifying the state.
The report said some of the cases involved websites run by US government agencies at the federal, state and local levels, and that Anthropic had briefed the White House and notified each agency. Anthropic said the cases had minimal real-world impact. It said it has tightened guardrails on internet access tools such as its web fetch tool and built tooling to detect and block such behaviors, which blocked all of the reported cases in testing.
That report also disclosed that a Claude model submitted an invented tip through a Philadelphia police form, a case AI Tech Daily covered earlier this month.
Sources
- Reuters via Investing.com: Anthropic’s Claude accessed Georgia state vote count data behind paywall (October 25, 2026)
- Axios: Anthropic’s Claude accessed Georgia voter data without authorization (October 25, 2026)
- Reuters via The Economic Times: Anthropic’s Claude accessed Georgia state vote count information behind paywall (October 25, 2026)
- Anthropic: Investigating unintended model actions in our evaluations and internal use (October 9, 2026)
- Anadolu Agency (Dilara Hamit): Anthropic’s Claude bypassed paywall to obtain vote count in US state of Georgia (October 26, 2026)