Topic analysis

U of T researchers demonstrate AI worm could target any online device

A team of University of Toronto researchers led by Nicolas Papernot has demonstrated a proof-of-concept AI-powered worm that uses freely available open-weight AI models to adapt its attack strategies as it spreads across interconnected devices, potentially targeting any online device at near-zero cost to hackers and evading current cybersecurity defenses. The researchers published their findings responsibly after consulting national security bodies, urging researchers, policymakers, and the public to take action—including practicing basic security hygiene like updating devices and using strong passwords—to develop countermeasures against this emerging threat.