Krux
February 26, 2026
300,000 ChatGPT Credentials Stolen as AI Speeds Up Hacking
Published: February 26, 2026 at 12:44 AM
Updated: February 26, 2026 at 12:44 AM
100-word summary
IBM's 2026 threat report reveals hackers stole over 300,000 ChatGPT credentials last year, treating AI tools like any other corporate software to breach. Attacks starting with public-facing apps jumped 44%, partly because AI helps attackers find security holes faster. Ransomware groups grew 49% as vulnerability hunting became the leading cause of breaches. The twist: companies are locking down their cloud apps but leaving AI platforms wide open with weak logins. The arms race is now symmetric: both sides are using AI to move faster, and whoever automates authentication wins.
What happened
IBM's 2026 threat report reveals hackers stole over 300,000 ChatGPT credentials last year, treating AI tools like any other corporate software to breach. Attacks starting with public-facing apps jumped 44%, partly because AI helps attackers find security holes faster. Ransomware groups grew 49% as vulnerability hunting became the leading cause of breaches. The twist: companies are locking down their cloud apps but leaving AI platforms wide open with weak logins.
Why it matters
The arms race is now symmetric: both sides are using AI to move faster, and whoever automates authentication wins.