Science & Tech · Cybersecurity · 1 hr ago
Google Says Hackers Are Using AI Agents to Speed Up Cyberattacks
Google says hackers are using autonomous AI agents to carry out several stages of cyberattacks with less human involvement.
The activity includes financially motivated attackers and suspected espionage groups, including actors linked to China.
The agents can analyze software patches and build exploits, helping attackers target vulnerabilities more quickly after fixes become public.
The basic methods, such as exploiting unpatched systems and using stolen credentials, are not new; AI helps attackers combine and scale them faster.
Some attackers are also stealing AI service credentials and cloud resources, then reselling access through underground services.
Google expects these tools to spread because open-source frameworks make them easier to use, and attackers can add new exploits.
Google says attackers are using autonomous AI agents to carry out multi-stage cyberattacks faster.
The frameworks can combine attack methods, exploit vulnerabilities and adapt to results with less human involvement.
Google says attackers are also using AI to study software patches and develop exploits for newly disclosed vulnerabilities.
The activity includes financially motivated actors and suspected espionage groups aligned with China.
Google says the tools can lower the barrier to entry and help scale attacks, but the underlying attack methods are not new.
- Who
- Threat actors, including financially motivated attackers and suspected China-aligned espionage actors.
- What
- They are using autonomous AI frameworks to automate parts of cyberattacks, including exploiting vulnerabilities and stealing credentials.
- When
- Google disclosed the findings on November 16, 2026. It said one attacker used a modified framework between late August and early October.
- Where
- Targets included organizations in the US aerospace sector and organizations across Asia. Google also reported attacks involving internet-facing servers.
- Why
- To speed up and scale intrusions, exploit vulnerabilities, and steal or resell access to AI credentials and cloud resources.
This story does not have two clearly opposing sides.
Google disclosed findings that threat actors are using autonomous AI frameworks to automate parts of cyberattacks.
Google observed attackers weaponizing vulnerabilities in PaperCut NG/MF after the flaws were disclosed.
An attacker used a modified PentAGI framework to scan organizations across Asia.
- Frameworks named
- ARTEX, CAIRN, CyberStrikeAI and PentAGI
- Observed in real attacks
- CAIRN and PentAGI
- US target sector
- Aerospace
- Asian scanning period
- Late August to early October
- Phishing platform
- Jieyou
3 sources
What people are saying
Adversary tradecraft has continued to transition beyond manual prompting to deploying autonomous multi-agent offensive frameworks—such as ARTEX, CAIRN, CyberStrikeAI, and PentAGI—to execute multi-stage intrusions at greater operational velocity.
The evolution of the Jieyou platform highlights a growing industry trend where PhaaS providers leverage LLMs as a modular force multiplier
Rather than manually editing payloads, low-skilled affiliates can now utilize natural language through the chatbot to dynamically customize pretext lures and phishing interfaces on the fly.