AI hackers speed up attacks as defenders struggle to keep pace
October 25, 2026
Summarised by Centrist
Artificial intelligence is allowing cybercriminals to launch attacks faster and on a greater scale, with security researchers warning that companies are struggling to keep pace.
The Wall Street Journal reports that hackers are increasingly using AI agents to automate intrusions, identify vulnerabilities and steal sensitive information.
Startup founder Colin Parsons discovered roughly $449 in unauthorised charges to his OpenAI and Anthropic accounts after credentials associated with an AI coding platform were apparently compromised.
Other users reported similar incidents.
Cybersecurity researchers say AI-driven hacking is no longer limited to sophisticated criminal organisations or state-backed groups. Less experienced attackers are also adopting the technology.
Microsoft recently warned that attackers were using a Chinese AI hacking tool called Artex against hundreds of companies. One reported attack compromised information belonging to approximately 68,000 South Korean banking customers.
Meanwhile, newly discovered security flaws potentially exposing as many as 70% of websites have prompted emergency patches and temporary shutdowns of government websites.
Researchers have also observed ransomware groups using AI to explore compromised networks and develop attack tools more rapidly.
“The pace that they operate at is faster than the average human,” a Sophos researcher said.
Security experts warn that AI is accelerating both the discovery of vulnerabilities and attempts to exploit them, creating an escalating contest between attackers and defenders.
Community Discussion
Comment Policy