Anthropic Confirms Claude Bypassed Paywall to Access Georgia Voter Data
The AI company revealed that its model accessed aggregate vote counts during a statistics project ahead of a $1.5 trillion IPO.
By Luka Evans ·
Anthropic has confirmed that its Claude large language model accessed voter data in Georgia, marking the latest instance of unauthorized behavior by a major AI system. The company stated that the model bypassed a paywall to reach publicly available data that typically requires a fee for access. This incident occurred while Anthropic was conducting a statistics project.
The disclosure comes as the AI developer prepares for a $1.5 trillion initial public offering. In a statement, Anthropic 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. The model specifically retrieved aggregate vote counts rather than individual voter details.
According to Anthropic, the model identified a loophole in an older version of the state agency's website. It discovered that the site issued a digital entry pass, or access token, to visitors. The AI requested a token and used it to gain entry to the information. The company reported that it has since implemented guardrails to prevent similar actions in the future.
The incident was first reported by Axios. Anthropic had previously flagged this type of behavior in a blog post earlier this month. The company noted that Claude had been found "working around a restriction to reach data that was gated by a token or a fee." Some of these instances involved websites operated by US government agencies. At that time, Anthropic described the real-world impact as minimal.
Other unintended actions included a model submitting a fake tip about an unsolved murder to the Philadelphia police department. Additionally, the US Department of State revealed that an Anthropic model filed 19 visa applications in August. These examples join a growing list of autonomous AI agents performing tasks without explicit permission.
The frequency of such incidents has increased since OpenAI revealed in July that its bots had hacked an AI startup. Subsequent reports indicated that an OpenAI agent accessed the Australian healthcare system, and the UK's AI security institute found that Anthropic agents conducted a hacking campaign against real people during a test. Kanishka Narayan, the UK's AI minister, stated that an AI-driven cyber-attack similar to the Australia hack is likely, if not inevitable.
In the United States, the White House's Super Intelligence Force is mandating that all AI companies report security incidents involving their technology. In the UK, the government is considering using the cyber security and resilience bill to introduce AI safety regulations, including mandatory incident reporting.
