AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: AI Attacks Without Humans: 5 Questions About Google’s Findings on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get audio and creator gear delivered free with Prime

  • Fast, free delivery on millions of items
  • Prime Video, Amazon Music and more included
  • Member-only deals all year
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

Google Threat Intelligence Group reported that it found an exposed server running software that used AI agents to identify vulnerabilities, write exploits and execute them without a person approving each step. Google describes this as one observed case; it forecasts that attacks with little or no human interaction may emerge more broadly in the coming months. The report does not identify the server’s operator or say how many sites were compromised.

Google’s Threat Intelligence Group says it found an exposed server running an AI-based framework that identified software vulnerabilities, drafted exploits and executed them without human approval at each step. In a report published November 16 with Mandiant, Google called this the first time it had seen attackers fully automate those tasks; it separately forecast that attacks requiring little or no human interaction could follow in the coming months.

The framework, which Google named exploit_agent, was hosted on an exposed staging server and connected to several commercial and locally run AI models. According to the report, it monitored widely used web software, compared code before and after security updates, inferred what weaknesses a patch fixed, then generated and ran attacks targeting those flaws. Google said this was an observed system, not a claim that fully autonomous attacks have become widespread.

Google also found victim data and scripts for later stages on the server, including tools for exploring networks and planting web backdoors. Data on the server suggested an interest in French websites in media, entertainment and education. A configuration file included comments in Russian. Google did not identify who operated the server, establish how many websites were breached, or say that every vulnerability the framework examined led to an attack.

The report also describes other AI-enabled activity. Google said it had seen open-source penetration-testing frameworks used in real intrusions, including CAIRN in operations targeting the U.S. aerospace sector and a customized PentAGI platform against organizations in Asia. These examples show AI tools being used in attacks, but they are distinct from the specific exploit_agent finding and do not establish that those operations were fully autonomous.

At a glance
reportWhen: Published November 16, 2026; broader au…
The developmentGoogle’s November 16 threat report describes an observed framework that automated vulnerability discovery and exploitation, while forecasting more autonomous attacks in the months ahead.

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

  1. 1Prompting a chatbothuman-driven
  2. 2AI built into malwareautomated runtime call
  3. 3AI automation in operationsAI in the loop, humans in charge
  4. 4Experimenting with autonomous frameworksexperimental autonomy

Now · in the wild

  1. 5AI tools run several attack stages“minimal human-in-the-loop”

Projected · forecast, not observed

  1. 6Simple attacks with little to no human interactionnear-term, 1–3 months
  2. 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.

Forecast

Solid: observed by Google. Dashed: projected by Google.

Google’s own label for today: stage 5 of 7. Fully automated exploitation exists in at least one case. Fully autonomous attacks as the norm are still a forecast.

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

Automation Changes Attack Speed

The finding matters because a system that can move from a software patch to a working exploit could shorten the time defenders have to update vulnerable systems. Automating repeated research and execution may also let an operator attempt more attacks with less direct effort. Google’s report documents a concrete case of that capability, even though it does not show how many victims were affected or how successful the framework was.

There is also a qualification in Google’s assessment: current AI-enabled tools can trade stealth for speed. Fast, high-volume activity may create patterns that security teams can detect. The practical concern is not that every attack now runs itself, but that automation can accelerate some stages of an intrusion and add pressure to patch, restrict access and monitor systems promptly.

Google recommends prioritizing actively exploited vulnerabilities, reducing the use of long-lived credentials, tightening administrator access and watching for activity bursts that are difficult to explain as ordinary human behavior. These are defensive measures; the report does not claim that any one measure would stop all AI-assisted attacks.

Products Worth Considering

Bose QuietComfort Headphones (2nd Gen), Wireless Noise-Cancelling Over-Ear Headphones, Eucalyptus Green

Bose QuietComfort Headphones (2nd Gen), Wireless Noise-Cancelling Over-Ear Headphones, Eucalyptus Green

LG 27GS60QC-B Ultragear 27-inch Curved Gaming Monitor (QHD, 180Hz, 1ms, 1000R)

LG 27GS60QC-B Ultragear 27-inch Curved Gaming Monitor (QHD, 180Hz, 1ms, 1000R)

Anker SOLIX C1000 Gen 2 Portable Power Station with 200W Solar Panel

Anker SOLIX C1000 Gen 2 Portable Power Station with 200W Solar Panel

As an affiliate, we earn on qualifying purchases.

One Observed Case, Two Forecasts

Google’s report, titled “GTIG AI Threat Tracker: Agentic Threat Activity Crosses a Significant Threshold,” is its quarterly review of AI use by criminal groups, state-backed actors and influence operations. The company said its findings draw on Mandiant incident-response work, tracking of threat actors and defenses on Google platforms.

The report places current activity at stage 5 of a seven-stage maturity scale, described as “minimal human-in-the-loop.” It 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, defined as three to six months. Those stages are forecasts, not events Google says have already become common.

This distinction explains why coverage of the report used different descriptions. The exploit_agent case supports saying that Google observed a fully automated vulnerability-discovery and exploitation process in at least one instance. The broader claim that fully autonomous attacks have arrived as a routine threat goes beyond the report: Google labels the present stage as involving minimal, rather than no, human input. Its October 30 discussion described a later stage in which a person sets the target and objective while agents carry out the attack chain.

““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

Who Ran the Server Remains Unknown

Google has not named the operator behind exploit_agent or said how many organizations were affected. The report describes data on the server as suggesting interest in French media, entertainment and education sites, but does not provide a confirmed victim count or establish the outcome of every attempted exploit. It also does not say whether the WordPress plugin access checks shown in the report led to a successful intrusion.

The scale of the capability remains uncertain. Google’s one observed framework is evidence that end-to-end automation can occur, but the report does not establish how often comparable systems are being used, how reliably they work, or whether operators can maintain stealth while attacks become more automated. Its timeline for broader autonomous attacks is a projection rather than confirmation that those stages will arrive on schedule.

Defenders Face a Near-Term Test

Google’s stated timeline points to little-to-no-human-interaction attacks within one to three months and fully autonomous attacks within three to six months. These are projections from the November report, not scheduled events or guarantees. The next evidence to watch for is whether threat researchers document more operations in which agents carry out most of an attack chain, and whether those systems become harder to detect.

For organizations, the immediate response is to review software exposure and patch actively exploited flaws, restrict administrative access, shorten the life of credentials where possible, and monitor for unusually rapid sequences of activity. Further reporting from Google or incident responders may clarify the operator, victims and effectiveness of exploit_agent. Until then, the confirmed development is a single documented case of automated vulnerability research and exploitation, alongside a warning that broader automation may follow.

Key Questions

What did Google say it observed?

Google said it found an exposed server hosting exploit_agent, a framework that compared software before and after updates, inferred vulnerabilities, drafted exploits and executed them without human approval at each step.

Does the report say fully autonomous cyberattacks are widespread?

No. Google described one observed case of automated vulnerability discovery and exploitation. Its report places current activity at “minimal human-in-the-loop” and forecasts broader autonomy in the future.

Who was targeted by exploit_agent?

Google said data on the server suggested interest in French websites connected to media, entertainment and education. It has not named the operator or provided a confirmed number of victims.

When does Google expect attacks with less human involvement?

Google projected attacks with little to no human interaction in one to three months, and fully autonomous attacks in three to six months, measured from its November 16, 2026 report. These are forecasts, not confirmed milestones.

What should organizations do now?

Google recommends prioritizing actively exploited software flaws, limiting long-lived credentials, tightening administrator access and monitoring for unusually fast bursts of activity.

Source: ThorstenMeyerAI.com

You May Also Like

Compilers for AI: Triton, XLA, and PyTorch 2.0 Inductor

Navigating the world of AI compilers like Triton, XLA, and PyTorch 2.0 Inductor reveals powerful tools that can transform your models, but there’s more to uncover.

What SemiAnalysis Says The 5X Means For AI Subscription Costs

SemiAnalysis estimates Claude subscriptions deliver about 5.4 to 5.6 times the API-priced usage of comparable ChatGPT plans on an agentic workload.

The Swarm Is The Weapon: Why Agentic Attacks Break The Defensive Playbook

Emerging autonomous AI collectives challenge existing cybersecurity defenses by operating in parallel, sharing knowledge instantly, and chaining vulnerabilities.

From Prototype To Practice: Understanding DeepSeek’s Coding Super-Agent

DeepSeek released an open-source prototype for an always-on coding agent, but says it is for research—not an official product launch.