Comcast Finds AI Is Supercharging Cyberattacks in Its 2026 Threat Report

PHILADELPHIA — Cyber attackers are using artificial intelligence to work faster, and the volume is staggering. Comcast Business’s 2026 Cybersecurity Threat Report, released Oct. 5, analyzed 79.3 billion security events detected across its customers over a 12-month period, or roughly 2,514 events every second.

The techniques were largely familiar. Phishing accounted for 25.4 billion events and drive-by compromise for 21.9 billion, the two most common attack methods in the data set. What changed, according to the report, is how quickly and cheaply attackers can run those campaigns at scale.

AI coding tools let attackers develop new exploits before software patches are available, narrowing the window between exposure and impact. Comcast says it blocks an average of 30 million threats a day across more than 1.2 billion connected devices.

Identity emerged as the central battleground. Non-human identities such as AI agents and application programming interfaces often carry broader permissions than an employee would receive, and the report recommends pairing identity management with continuous behavioral analysis.

“Attackers have stopped needing to break in,” said Noopur Davis, chief information security and product privacy officer at Comcast, in coverage of the report by Telecompetitor. She explained that intruders often log in with stolen credentials and then move through a network the way an employee would.

Other findings included 288.9 million active scans hunting for vulnerable devices, 5.8 million attempts to gain covert access to employee browsers through malicious extensions or session theft, and 57,000 distributed denial-of-service events driven by botnets.

The report is the fourth annual edition from the Philadelphia-based company. For local businesses, the takeaway is that familiar defenses still matter, but they now have to operate at machine speed to keep up.