Osegi
Security·1 min read

Google Threat Intelligence Group sees fully automated AI attacks

Google's threat researchers say some attackers now let AI run IT attacks from start to finish, with no human steering each step, according to their latest report.

A dark server room with glowing code on a screen, symbolising automated AI-driven cyberattacks.
Image: Security

Cyberattacks that run themselves are no longer a thought experiment. Google's Threat Intelligence Group, the company's team that tracks hackers and their tools, says in its latest report that it has seen attackers who automate IT attacks completely with the help of AI.

That is the core of the finding: not just AI as a writing aid for phishing emails or a helper for debugging malware, but AI doing the work of an attack end to end. The report describes these attackers and how they use the technology.

What does fully automated mean here?

In plain terms, the steps an attacker normally performs by hand are handed over to AI systems. Fewer people are needed, and an attack can in principle be repeated against many targets at once. The details of which tools and models are involved are in Google's report itself, and they are not spelled out here.

Why it matters for everyday users

The Threat Intelligence Group watches real attacker activity, so this is an observation from the field rather than a lab warning. For companies and ordinary users alike, the practical advice stays familiar: keep software updated, use multi-factor authentication and treat unexpected messages with suspicion. Automation does not change those basics. It mainly means attackers can try more, faster.

Apps worth a look

SUMO: the number puzzle

Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.

Tags:googlethreat intelligencecyberattacksaiautomationsecurity

More from this day

See all →

More in: Security

See all →