Google Threat Intelligence Group (GTIG) says threat actors may be using large language models and other AI tools to analyze patches, product versions, vulnerability notices and proof-of-concept code faster—and potentially weaponize already disclosed vulnerabilities. That is a warning about a possible shift in attackers’ efficiency, not proof that AI caused the rise in exploitation GTIG measured. Its data show more observed exploited vulnerabilities in 2026, while zero-day exploitation rose more modestly.
What Google is warning about
In a post published September 30, 2026, GTIG says it is possible that threat actors are using LLMs and other AI tools to automate comparisons between software versions, patches, vulnerability announcements and proof-of-concept code. That could help them turn disclosed flaws into working attacks against organizations that have not yet applied a fix.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
Network Security, Firewalls, and VPNs | $66.62 | Buy on Amazon |
| 2 |
|
Network Security, Firewalls, and VPNs: . (Issa) | $57.99 | Buy on Amazon |
| 3 |
|
TP-Link ER605, Wired Gigabit VPN Router | $49.99 | Buy on Amazon |
| 4 |
|
Cybersecurity for Small Networks: A Guide for the Reasonably Paranoid | $33.89 | Buy on Amazon |
The distinction matters: this is a hypothesis about how attackers may be using AI, not a finding that AI has been proven to cause the increase in exploitation. GTIG’s report, credited to Robin Grunewald, Supriya Mazumdar and Kelli Vanderlee, covers vulnerability disclosures from January 1, 2025 through August 31, 2026. Read GTIG’s analysis, “Vulnerability Discovery and Exploitation Trends in the AI Era.”
What GTIG’s 2026 figures show
GTIG reports that both monthly vulnerability disclosures and its observed count of exploited vulnerabilities increased in 2026. The figures describe activity tracked by the group; they are not a count of every vulnerability or attempted attack worldwide.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →#1 Best Overall
| Measure | 2025 | 2026 period reported |
|---|---|---|
| Monthly vulnerability disclosures | not stated as a yearly average (GTIG) | 5,045 in January and 10,740 in August (GTIG) |
| Observed exploited vulnerabilities | Average of 10.5 per month (GTIG) | Average of 18 per month, January–August (GTIG) |
| Observed zero-day exploitation | Average of 8 per month (GTIG) | Average of 11 per month, January–August; 22 in August (GTIG) |
Zero-days are vulnerabilities exploited before a fix is publicly available. GTIG says they made up 62% of its observed exploited vulnerabilities during January–August 2026, even though zero-days were a small share of all disclosed vulnerabilities. That percentage applies to the exploited vulnerabilities GTIG observed in that period; it is not the share of all disclosed flaws that were attacked.
Why more CVEs do not automatically mean more danger
A higher disclosure count is not, by itself, evidence that every added entry is exploitable, severe or under attack. GTIG cautions that automated CVE Numbering Authority assignment policies can inflate raw totals. As an example, it cites roughly 5,000 CVEs whose descriptions included “Linux Kernel” from January through August 2026, with no observed in-the-wild zero-days in that group.
Rank #2
- Available with the Cloud Labs which provide a hands-on, immersive mock IT infrastructure enabling students to test their skills with realistic security scenarios
- New Chapter on detailing network topologies
- The Table of Contents has been fully restructured to offer a more logical sequencing of subject matter
- Introduces the basics of network security—exploring the details of firewall security and how VPNs operate
- Increased coverage on device implantation and configuration
GTIG also distinguishes its own vulnerability risk ratings from CVSS severity scores. A disclosure total, a severity score and evidence of active exploitation answer different questions; none should be treated as a direct substitute for the others.
AI-assisted discovery and the race to patch
GTIG describes a possible collision between faster vulnerability discovery and faster exploitation. Its example is CVE-2026-1731, an unauthenticated OS command-injection flaw in BeyondTrust Privileged Remote Access and Remote Support. According to GTIG, third-party research agent Hacktron AI discovered the flaw autonomously. GTIG then observed a threat cluster exploiting it within four days of public disclosure, followed by five additional clusters within seven days.
Recommended Free Tools
Rank #3
- 【Five Gigabit Ports】1 Gigabit WAN Port plus 2 Gigabit WAN/LAN Ports plus 2 Gigabit LAN Port. Up to 3 WAN ports optimize bandwidth usage through one device.
- 【One USB WAN Port】Mobile broadband via 4G/3G modem is supported for WAN backup by connecting to the USB port. For complete list of compatible 4G/3G modems, please visit TP-Link website.
- 【Abundant Security Features】Advanced firewall policies, DoS defense, IP/MAC/URL filtering, speed test and more security functions protect your network and data.
- 【Highly Secure VPN】Supports up to 20× LAN-to-LAN IPsec, 16× OpenVPN, 16× L2TP, and 16× PPTP VPN connections.
- Security - SPI Firewall, VPN Pass through, FTP/H.323/PPTP/SIP/IPsec ALG, DoS Defence, Ping of Death and Local Management. Standards and Protocols IEEE 802.3, 802.3u, 802.3ab, IEEE 802.3x, IEEE 802.1q
GTIG says the campaigns used the vulnerability for initial access and then involved actions including privilege escalation, data exfiltration and delivery of secondary payloads. This example shows why time to remediation matters, but it does not establish that AI was used by the attackers in those campaigns.
The group calls the higher-risk profile it sees in some AI-discovered vulnerabilities an early indicator, not an established trend. Its summary says AI-assisted discovery found proportionally fewer low-risk and more moderate-risk vulnerabilities, as well as more flaws leading to remote code execution. That does not mean every AI-discovered flaw is severe, or that AI alone explains those characteristics.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What organizations should do
GTIG recommends moving away from unprioritized mass-patching toward threat-intelligence-driven triage, targeted edge defense and automated, agentic remediation. In practice, this means keeping patch management in place while directing the fastest response to vulnerabilities that combine credible exploitation evidence with exposure in the organization’s environment.
- Prioritize evidence, not just volume. Use reports of active exploitation and relevant threat intelligence alongside severity ratings and asset exposure when deciding what to fix first.
- Find exposed systems quickly. Identify internet-facing products and services that contain affected software, especially where a flaw could provide initial access.
- Speed up safe remediation. Automate asset identification, patch deployment and verification where the organization can do so reliably; retain controls for testing and exceptions.
- Keep patching broadly. Risk-based triage helps allocate urgency, but it is not a reason to ignore less-publicized vulnerabilities or leave a backlog unmanaged.
The practical concern is not that every new disclosure will become an AI-powered attack. It is that attackers may be able to extract useful exploit details from public information more quickly, making delays in identifying and fixing exposed systems more consequential.
Free tools Windows power users keep installed
One-click scans. No signup required.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
