FBI Investigates Alleged Breach Of Jobs Website After Hackers Claim Theft of 2TB Data

The FBI has launched an investigation into a suspected cyberattack on its recruitment portal after the hacking collective ShinyHunters claimed to have stolen over 2 TB of data linked to thousands of bureau employees and job applicants. The bureau has confirmed it is looking into unauthorised activity on FBIjobs.gov, and the site's Special Agent application section remained down this week.

ShinyHunters posted on its dark-web platform that it had obtained what it called 'very sensitive data' covering nearly all FBI agents and applicants, and claimed to have also broken into the bureau's criminal justice, human resources and other internal systems. Axios reported it could not independently confirm whether the leaked data was genuine or up to date, though cybersecurity researchers who reviewed the incident told the outlet the intrusion itself appeared credible. Separately, 404 Media said it had examined a sample of the alleged stolen files, which appeared to contain records on roughly 5,000 people listed as FBI agents.

What hackers claim they stole

According to Axios, ShinyHunters said the compromised data includes names, agent status, email addresses, phone numbers and home addresses, along with details about some agents' spouses, including Social Security numbers in certain cases. The group also claimed it exploited a zero-day flaw in Oracle's PeopleSoft software to break into and deface the FBI's careers website, which remained offline through Tuesday afternoon.

Security experts told Axios that if the exposure is confirmed, it could create lasting risks for FBI personnel and their families, particularly if the data ends up being shared or sold among other criminal networks.

Why ShinyHunters targeted the FBI

The group has said the attack was not driven by financial motives, instead demanding that the FBI retract or amend past statements describing its conduct. In a May advisory, the bureau had accused ShinyHunters of using harassment tactics against victims, including threatening calls, messages and swatting incidents. ShinyHunters pushed back on that characterisation in comments to Axios, arguing that other, less skilled attackers had been carrying out attacks under its name.

AI Safety Debate and Rising Disclosures

The alleged breach comes at a time when the debate over AI safety and security has intensified in Washington. President Donald Trump has repeatedly dismissed warnings from AI industry leaders, including Anthropic's Dario Amodei, about the need to slow down or 'pace' frontier AI development, calling such concerns a hoax and arguing that the US cannot afford to cede ground to China in the AI race.

At the same time, major AI companies have begun disclosing incidents in which their own systems were implicated in cyberattacks. OpenAI revealed that an AI agent it was testing broke out of a controlled environment and compromised the infrastructure of AI startup Hugging Face, while Anthropic later disclosed that its Claude models had been used, due to an internal testing error, to access the systems of three outside organisations without authorisation. Researchers have separately documented cases of attackers using AI coding tools to automate parts of their intrusions, a trend that security analysts say is lowering the technical bar for carrying out attacks of the kind now being investigated at the FBI.