Back to blog

Blog Post

Beacon: Google Warns that Fully-Automated AI Attacks are Coming Soon

Share on social

Nov 19, 2026

Lorem ipsum

Table of contents

Share on social

Join the newsletter
Beacon: Google Warns that Fully-Automated AI Attacks are Coming Soon

This story appeared in our weekly cybersecurity newsletter Beacon. Sign up to get weekly updates on the latest news, findings and insights, straight to your inbox every Thursday at 10AM.

Top Story This Week:

Google Warns that Fully-Automated AI Attacks are Coming Soon

Google Threat Intelligence Group (GTIG) says attackers are now using agentic workflows built on AI-based offensive frameworks to automate vulnerability discovery and exploitation, speeding attacks up to machine speed. Its quarterly AI Threat Tracker, published on November 16, projects that within three months attackers will launch unsophisticated attacks needing little to no human interaction, relying mostly on multiple AI agents. Within six months, it expects sophisticated, fully autonomous agentic attacks to emerge.

GTIG says attackers are systematically abusing open-source penetration testing suites, including ARTEX, CAIRN, CyberStrikeAI and HexStrike AI. In one example, a multi-agent system called exploit_agent fully automated finding and exploiting vulnerabilities in attacks on French media, entertainment and education websites. GTIG says that in the past quarter it identified, for the first time, threat actors using LLM frameworks to automate discovery and exploitation with no human in the loop.

Why it Matters

Google’s warning fits what the data already shows: AI isn’t creating new kinds of attack so much as making existing ones faster, cheaper and more scalable. The estimate for mean time to exploit is now negative seven days, meaning exploitation often begins before a patch is even released.

At the same time, defenders are being buried: disclosure volumes are surging, only 26% of CISA KEV entries were fully remediated in 2025, and the average organization can see only around 62% of its external attack surface. Autonomous attackers will be quickest to find the assets nobody is watching and the exposures nobody has validated. The old assumption of a comfortable patch window no longer holds.

As our AI vs. the Patch Window research argues, defense has to shift to continuously finding and closing exposures before attackers can reach them.

Advice

Assume you are exposed, and focus on closing exposures before attackers’ agents find them. Preemption means acting on what could be exploited, not waiting for a CVE and a scheduled patch cycle to start the clock.

Make attack surface visibility continuous and comprehensive. A point-in-time scan is out of date within hours, so keep discovering assets across cloud, identity and on-premises, including unmanaged ones such as acquired subsidiaries’ infrastructure, forgotten development environments and marketing SaaS deployments.

Validate by exploitability, not theoretical severity. Test whether an exposure can actually be exploited in your environment, so teams work from a short, trusted list of proven problems rather than a long queue of theoretical ones.

Know the attacker reality. Overlay your validated exposures with intelligence on what attackers are targeting now, including active campaigns, dark web and forum activity, and the AI tools in use, so you fix what is actually in their sights first.

Send high-confidence findings straight to remediation. With exploitation starting within hours of a patch, a proven, actively targeted exposure shouldn’t wait in a queue.

Discover More

How AI Is Collapsing Exploitation Timelines

Defenders used to have one advantage…time. A vulnerability would be discovered, a disclosure or advisory would follow, and security teams would have sometimes months to patch or mitigate before attackers caught up. AI is changing the economics of vulnerability discovery, exploit development, and attack execution - making them accessible to a much broader range of threat actors. Read more.

Preemptive Threat Exposure Management in the Age of AI

With the introduction of Mythos and other frontier AI models, there has been a flood of media coverage and hype about what these technologies could mean for the security industry. This has left many organizations trying to understand what it means for them and how they can best prepare for the impact of these models. Read more.

Weekly News Digest

Operation TorNado Leads to More than 350 Arrests Across 14 Countries

An international law enforcement operation, known as Operation TorNado, has resulted in the arrest of 354 individuals across 14 countries for their involvement in illegal drug trafficking through dark web marketplaces. Led by the FBI alongside international law enforcement agencies, the operation also saw the seizure of more than 5.7 tonnes of illegal drugs, over $234 million in cash and cryptocurrency, and 209 firearms. Running from May 2025 to October 2026, the operation targeted drug suppliers, buyers, and trafficking networks, including those involved in the distribution of fentanyl and other dangerous substances. Despite the significant number of arrests, investigations remain ongoing as authorities continue efforts to identify and dismantle further criminal networks operating online.

Russian Cyberattackers Hit Expedia - Claiming Nearly A Billion Traveler Records Exposed  

Expedia Group has confirmed it was targeted in a major cyberattack after a Russian hacking group claimed to have stolen almost one billion customer records. The breach, reportedly carried out through a compromised third-party data provider, could have exposed millions of travellers personal details, including email addresses, loyalty programme information, account credentials, and travel profiles. The ransomware group, known as n0n, is threatening to release the stolen information unless a settlement is reached. Expedia says the full extent of the breach is still under investigation and that it has taken immediate steps to contain the incident, reassuring customers that its platform remains safe to use.

Check Point Gateways Hacked via PHP2Shell to Deploy VMware ESXi Wiper

Hackers initially targeted Israeli organizations on October 30th, before a second wave of exploitation on November 2nd affected companies in multiple countries and sectors. A single actor is believed to have conducted the entire campaign, from initial access to launching PHP web shells. In addition to deploying destructive data-wiping malware, in some instances researchers found political messages left on impacted hosts that claimed to be from “Dream Security Ltd” and explicitly rejected ransom payment negotiations.  

‍

Lizzie Clark

Author

Lizzie Clark

Marketing Executive at Searchlight Cyber

Lizzie is an experienced IT and cybersecurity marketing professional with six years of specialist experience in the industry. Lizzie produces a range of content - from blogs and long-form articles to newsletters and social media - with a focus on writing that informs and engages technical audiences.

Related Blog Posts

November 18, 2026

AI Is Changing Vulnerability Discovery. But Can Security Keep Up?

November 13, 2026

Third Party Sprawl: How to Manage a Constantly Expanding Attack Surface

November 12, 2026

Beacon: Hundreds of Artex AI Hacking Tools Found Exposed Online

November 9, 2026

New Indicator of Compromise hunting tool added to Searchlight Investigate

November 6, 2026

Does a Ransomware Leak Site Listing Confirm a Breach?

November 4, 2026

CVE Intelligence now accessible across all Searchlight tools

Never miss a beat

Get all news and updates about Searchlight Cyber, directly in your inbox.

Subscribe
Please enter a valid email address.
Background Gradient