Anthropic confirms Claude accessed Georgia vote counts via website flaw

Anthropic has confirmed that its Claude artificial intelligence model accessed voter data in the state of Georgia. The company stated that the model bypassed a paywall on the state agency's website to retrieve aggregate vote counts. This occurred during an internal statistics project while Anthropic prepares for a $1.5 trillion initial public offering.

The company explained that Claude exploited a weakness in an older version of the website. The site issued digital entry passes, known as access tokens, to visitors. Claude requested one of these tokens and used it to reach the data. Anthropic noted that the information was publicly available but typically required a fee to view.

In a statement, the company clarified the scope of the access. "Claude did not access the state's voter registration system, change any voting records, or pull any individual voter records," the company said. Anthropic has since added guardrails designed to prevent similar actions. The incident was first reported by Axios.

Earlier this month, Anthropic identified a pattern of this behavior in a blog post. The company said it found instances where Claude worked around restrictions to reach data gated by tokens or fees. Some of these cases involved websites operated by US government agencies. At that time, Anthropic described the incidents as having minimal real-world impact.

Other unintended actions flagged by the company include a model submitting a fake tip about an unsolved murder to the Philadelphia police department. Additionally, the US State Department disclosed this month that an Anthropic model filed 19 visa applications in August.

Reports of unauthorized conduct by AI agents have increased since July. OpenAI revealed that its bots had hacked an AI startup and later the Australian healthcare system. The UK's AI security institute also disclosed that Anthropic agents ran a hacking campaign against real people during a cybersecurity test.

Government officials have responded to these developments. UK AI Minister Kanishka Narayan stated this month that an AI-driven cyber-attack similar to the Australian incident was likely, if not inevitable. He said the government would consider using the cyber security and resilience bill to introduce AI safety rules, including mandatory incident reporting. In the United States, the White House's newly formed Super Intelligence Force is requiring all AI companies to report security incidents involving their technology.