🔍 Read the full analysis: A Closer Look At Google’s Claims About AI Attacks Without Humans on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Google’s Threat Intelligence Group says it found an exposed server running software that automatically analyzed security updates, developed exploits and executed them without a person approving each step. The finding is one observed case; Google’s forecast that attacks with little or no human involvement will become more common in coming months is not evidence that such attacks are already routine.
Google’s Threat Intelligence Group says it found an exposed server running a framework that analyzed software updates, developed exploits and launched them without a person approving each step—a finding the company describes as its first observed case of fully automated vulnerability discovery and exploitation. The November 16 report documents one such system, while its prediction that attacks with little or no human involvement will expand in the coming months remains a forecast, not a description of established practice.
The report, produced by Google Threat Intelligence Group (GTIG) with Mandiant, calls the framework “exploit_agent.” Google says it was hosted on an exposed staging server and connected to multiple commercial and locally run large language models. The system compared software code before and after security updates, inferred which vulnerabilities a patch addressed, drafted attacks and automatically executed them.
Google also says the server contained stolen victim data and scripts for later stages, including searching within networks and installing web backdoors. The victim data suggested the operator was interested in French websites in media, entertainment and education. Google did not identify the operator or say how many websites were breached. The report includes an example of the framework examining access checks in a WordPress plugin, but does not say whether that examination resulted in an attack.
GTIG says it has also observed open-source AI-assisted penetration-testing frameworks used in real intrusions. It names CAIRN, which it says was used against targets including the US aerospace sector, and a customized PentAGI platform used against organizations in Asia. These examples show AI tools appearing in cyber operations, but they are separate from the specific claim that exploit_agent automated vulnerability discovery and exploitation end to end.
Google threat report · Nov 16, 2026
How far AI attacks have come, by Google’s count
Google sorts attacker use of AI into seven stages. Four are history, one is now, two are a forecast.
Google’s seven stages
Observed · end of 2025 – Sep 2026
- 1Prompting a chatbothuman-driven
- 2AI built into malwareautomated runtime call
- 3AI automation in operationsAI in the loop, humans in charge
- 4Experimenting with autonomous frameworksexperimental autonomy
Now · in the wild
- 5AI tools run several attack stages“minimal human-in-the-loop”
Projected · forecast, not observed
- 6Simple attacks with little to no human interactionnear-term, 1–3 months
- 7Sophisticated, fully autonomous attacksmedium-term, 3–6 months
Two headlines, one report
“Have arrived”
The Decoder, Nov 16
Follows one observed case: a framework that found and exploited flaws with no human in the loop.
Evidence · one server“Coming soon”
BleepingComputer, Nov 16
Follows the timeline: stages 6 and 7 are Google’s forecast.
ForecastSolid: observed by Google. Dashed: projected by Google.
Sources: Google Threat Intelligence Group, “GTIG AI Threat Tracker”, Nov 16, 2026 (Figure 2, LLM maturity timeline); GTIG and Mandiant, Oct 30, 2026; BleepingComputer and The Decoder, Nov 16, 2026. Stage names shortened by us.
Graphic: Powered by Thorsten Meyer AI
What Automated Exploitation Changes
Automating the work of identifying and exploiting a newly patched flaw could shorten the time between a security update becoming public and attackers attempting to use the weakness it fixed. That makes timely patching and monitoring relevant even for organizations that have not seen evidence of this particular framework. Google’s report does not establish how many attacks exploit_agent carried out or how widespread this capability is, so the finding should not be read as proof that automated attacks are now the norm.
Google says current tools can trade stealth for speed: rapid, machine-generated activity may be easier to spot because it is noisy. The report recommends prioritizing actively exploited vulnerabilities, limiting long-lived credentials, restricting administrator access and looking for bursts of activity that do not resemble normal human behavior. These are defensive steps based on the reported risk, not a guarantee that an organization will prevent every attack.
Popular right now

Hisense E7 Cinema Series 55-Inch Hi-QLED Mini-LED 4K UHD Smart Fire TV

Ninja CREAMi Scoop & Swirl 13-in-1 Ice Cream and Soft Serve Maker, NC701

Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation
As an affiliate, we earn on qualifying purchases.
One Case, Two Timelines
GTIG’s report, titled “GTIG AI Threat Tracker: Agentic Threat Activity Crosses a Significant Threshold,” is a quarterly account of how criminal groups, state-backed actors and influence operations use AI. Google says its findings draw on Mandiant incident-response work, threat-actor tracking and platform defenses. The exploit_agent server is the report’s central example of end-to-end automated exploitation; it is not presented as a count of similar systems or victims.
Google separates that observed incident from a seven-stage maturity timeline. In its account, stage five—“minimal human-in-the-loop”—is the current stage. The company projects attacks with little to no human interaction in the near term, defined as one to three months, and fully autonomous attacks in the medium term, three to six months. In an October 30 post, GTIG described a later stage in which people set a target and objective while multi-agent systems carry out the attack chain. Those later stages are projections, not findings that the report says have already become standard.
That distinction helps explain differing headlines about the report. Coverage describing automated attacks as having “arrived” reflects Google’s account of the exploit_agent case. Coverage saying they are “coming soon” reflects the company’s forecast of broader, more autonomous operations. Both refer to different claims: one observed framework and a predicted shift in attacker behavior.
““In the past quarter, GTIG identified threat actors using LLM frameworks to fully automate vulnerability discovery and exploitation without a human in the loop for the first time.””
— Google Threat Intelligence Group
The Scope of the Finding
Google has not disclosed who operated exploit_agent, how many sites were compromised or when the activity occurred. Its report says victim data pointed to French media, entertainment and education websites, but does not give a confirmed victim count or establish that the WordPress plugin analysis shown in the report led to a successful intrusion.
The report also does not establish how often fully automated exploitation is occurring across the wider threat landscape. Google’s timeline is an assessment of likely development, and its projections for the next one to six months are not confirmed outcomes. The extent to which attackers can keep such systems effective while avoiding detection remains uncertain.
What Defenders Should Watch
Google’s report gives no public date for additional findings about exploit_agent or its operator. Its near-term forecast puts attacks with little to no human interaction within one to three months, followed by a projection of fully autonomous attacks within three to six months. Those windows describe Google’s expectations, not scheduled events or guarantees.
For security teams, the immediate response is to review patching priorities, especially for vulnerabilities known to be actively exploited, and check controls around administrator access and long-lived credentials. Monitoring for unusually rapid, high-volume activity may help identify automated operations, though the report does not specify a single detection method that will work in every environment. Further information about the framework, its victims and whether similar systems are spreading has not been provided.
Key Questions
What did Google say it found?
Google says it found an exposed server hosting exploit_agent, a framework that analyzed code changes after software updates, developed exploits and executed them automatically. The company described it as its first observed case of fully automated vulnerability discovery and exploitation without a human in the loop.
Does the report show that fully autonomous attacks are now common?
No. Google describes one observed framework that automated vulnerability discovery and exploitation. It labels the current broader stage as involving minimal human involvement and forecasts more autonomous attacks in the future; it does not say that fully autonomous operations are already routine.
Who was targeted by exploit_agent?
Google says data on the server suggested interest in French websites associated with media, entertainment and education. The report does not identify the operator or confirm how many websites were compromised.
What should organizations do in response?
Google recommends prioritizing patches for actively exploited vulnerabilities, limiting long-lived keys and passwords, restricting administrator access and watching for bursts of activity that may indicate automated behavior. These measures reduce exposure but are not a guarantee against attacks.
Source: ThorstenMeyerAI.com
