AI is lowering the cost of hacking

Anthropic's latest threat intelligence report finds that AI agents with advanced cyber skills are making attacks so easy that hackers no longer need to target only high-value companies. Smaller firms are now tempting targets.

“why do robbers rob banks? Because that's where the money is.”

Attackers using Claude to breach systems

Anthropic found attackers using Claude to navigate corporate systems they barely understood, identify valuable data, and write code to exploit vulnerabilities. In one case, an attacker with a stolen developer token took full administrative control of a company's cloud environment in roughly three hours.

ShinyHunters breach shows AI's scale

After breaching a software provider, attackers Anthropic described as suspected affiliates of ShinyHunters extracted data belonging to roughly two hundred of its customers. Anthropic said AI agents performed nearly all the work in the intrusion.

Vibe hacking automates intrusions

During the software-provider intrusion, attackers extracted more than two thousand one hundred sets of Azure AD authentication tokens across more than forty corporate cloud environments in about thirty four hours. Anthropic calls the technique of letting AI try different tactics until one succeeds vibe hacking.

Russian attackers target AI companies

A Russian-speaking attacker stole roughly twenty six gigabytes of data from one victim and sought one point five million to two point five million dollars through extortion or dark web sales. A later campaign from the same infrastructure targeted roughly thirty AI companies in about four days.

Scammers use AI for mass catfishing

A China-based app studio used Claude to operate more than four thousand seven hundred personas across a network of dating apps, conversing with at least twenty five thousand people over two weeks in April. Real gig workers handled tasks like live video calls to make the service seem authentic.