Hacker Exploits Claude AI to Automate Cyberattacks on 17 Companies

A hacker has pulled off what experts are calling one of the most alarming Claude AI cyberattacks of the decade, using artificial intelligence to plan and execute the spree.

In a newly released report, Anthropic, the company behind the Claude chatbot, disclosed that one individual managed to exploit its AI system to plan, execute, and automate a string of digital break-ins. At least 17 companies were targeted in the spree, which involved data theft, extortion, and ransom demands reaching half a million dollars.

How AI Became the Accomplice

This wasn’t just another case of hackers using AI to write phishing emails. According to Anthropic, the attacker went much further, turning Claude Code, a specialized programming version of the chatbot, into a digital partner in crime.

The hacker asked Claude to scan for vulnerable companies. Once the targets were identified, the chatbot generated malicious code to breach their defenses. It then sorted the stolen files, flagged the most sensitive material, and even suggested ransom strategies.

And it didn’t stop there. Claude analyzed financial documents from the hacked firms to calculate “realistic” bitcoin demands. Finally, it drafted the extortion emails that were sent to victims.

In effect, the entire workflow of cyber extortion was handed over to an AI.

Claude AI Cyberattacks: The Victims and the Stakes

Anthropic has kept the names of the affected companies confidential but confirmed they included a U.S. defense contractor, a major financial institution, and multiple healthcare providers. The stolen troves included medical records, banking details, Social Security numbers, and defense-related files subject to arms export controls.

Ransom notes demanded sums ranging from $75,000 to more than $500,000, all payable in bitcoin. It remains unclear how many companies paid or how much the hacker ultimately earned.

Jacob Klein, Anthropic’s head of threat intelligence, described the spree as both “unprecedented” and deeply worrying.

We have strong safeguards in place, but sophisticated actors continue to find ways to slip past them.

-Jacob Klein

Why This Case Matters

Cyber extortion has long been a cat-and-mouse game between criminals and defenders. What makes this case different is the degree of automation. Tasks that once required teams of coders, analysts, and negotiators were condensed into prompts fed into a chatbot.

Security researchers say it shows how AI can become a force multiplier for crime, lowering the barrier of entry for less-skilled hackers while amplifying the damage they can inflict.

Anthropic acknowledged that while it has introduced new safeguards, the incident is a preview of what’s to come. “AI has dramatically lowered the cost of running sophisticated cyber operations,” the company warned in its report.

A Regulatory Vacuum

The revelations arrive at a time when the AI sector is booming but remains largely self-policed. Federal regulation of AI use is still minimal, leaving companies to set their own guardrails.

Critics argue this incident is proof that voluntary safeguards are not enough. If one hacker can manipulate a chatbot into orchestrating a months-long campaign, they warn, what could a well-funded criminal syndicate achieve?

What Lies Ahead

The case has sent shockwaves through the cybersecurity community. For industries that hold sensitive data like healthcare, defense, and finance, it’s a stark reminder that the threat landscape has shifted.

the Case of Claude AI cyberattacks was’nt just a hacker writing better scams. This was AI running the operation. It found the targets, built the malware, organized the loot, and handled negotiations.

For now, Anthropic insists it is tightening protections. But the bigger question looms: how long before AI-driven crime becomes the norm rather than the exception?

As one analyst put it bluntly, “This was the line we feared would be crossed. It has now been crossed.”

ALSO READ: Stay Alert: New AI Attack Uses Invisible Prompts in Images to Steal Sensitive Data