China-Linked Hackers Used Anthropic AI To Automate Attacks

Benzinga

Anthropic said China-linked hackers used its AI tools to automate up to 90% of a September cyberattack targeting about 30 companies and foreign governments. The attackers posed as security auditors, enabling phishing, scanning, and data extraction before Anthropic shut down their accounts and tightened safeguards.

Transcript

00:00It's Benzinga, bringing Wall Street to Main Street.

00:02China-linked hackers used Anthropic's AI tools to automate a September hacking campaign

00:07targeting roughly 30 corporations and foreign governments, Anthropic said Thursday.

00:12Investigators said 80% to 90% of the operation ran automatically, as hackers stitched together

00:18tasks like scanning, phishing, and data extraction into near-autonomous workflows.

00:24Anthropic blocked the accounts but confirmed several successful intrusions that stole

00:28sensitive information, though not from the U.S. government.

00:31The attackers bypassed safeguards by posing as security auditors and directing Claude

00:36to perform internal queries.

00:38The company said it has updated its misuse detection systems as U.S. officials warn of

00:42growing state-backed efforts to weaponize AI tools.

00:45For all things money, visit Benzinga.com.

Category

Transcript

Be the first to comment

Recommended