Table of Contents
Artificial intelligence is quickly becoming a major tool in the world of cybercrime, and a new report from Google suggests things are starting to take it much more seriously. According to the company’s latest findings, hackers are now using AI tools to help discover zero-day vulnerabilities, including one that could bypass two-factor authentication protections.
The report comes from Google’s Threat Intelligence Group, which recently shared details about how cybercriminals are increasingly relying on AI to improve attacks, create malware, automate phishing campaigns, and even uncover software flaws that developers never knew existed.
Hackers seem to be using AI-generated code to bypass 2FA
One of the most concerning discoveries involved hackers attempting to exploit a previously unknown vulnerability in a popular open-source web-based system administration tool. The flaw could reportedly allow attackers to bypass 2FA protections, meaning they could potentially access accounts with just a password and skip the extra security code normally sent to your phone or email.
Latest tech deals.
- AMD Ryzen™ 9 9900X 12-Core - was $499 now $331
- Sony INZONE M10S 27-inch - was $899 now $748
- Thermaltake LCGS View i1470S-170 - was $2,399 now $2,099
- Apple 2026 MacBook Air - was $1,999 now $1,891
- Bose QuietComfort Headphones - was $359 now $169
- LG 77-inch OLED C6 - was $2,499 now $1,999
- MSI Prestige 16 Flip - was $2,149 now $1,870
- Hisense 85-inch U6 Pro Series - was $1,999 now $997
Prices correct as of November 5th, 2026.
Google believes AI played a major role in helping the attackers discover and weaponize the flaw. The company noticed several signs that strongly pointed toward AI-generated code. The Python script used in the attack reportedly included unusually detailed educational comments, overly structured formatting, and even a fake or “hallucinated” CVSS security score, which is something commonly seen in content produced by large language models.
Fortunately, Google says it worked with the affected vendor before the vulnerability could be widely exploited. The company believes the hackers were planning a large-scale attack campaign, but the issue was patched before that could happen.
Zero-day exploit successfully found with help from AI
What makes this situation especially worrying is that this appears to be the first confirmed case where AI was successfully used to help develop a zero-day exploit. Security researchers have warned for years that AI could eventually become a major cybercrime weapon, and Google now says that moment may already be here.
The report also explains that AI is being used in many other dangerous ways. Some malware strains can now dynamically modify their own code while running in order to avoid detection from antivirus software. Other malicious tools can automatically generate fake code, create exploit payloads on the fly, or add layers of obfuscation that make attacks harder to analyze.
Google also noted that state-sponsored hacking groups from countries like China and North Korea are heavily experimenting with AI tools for vulnerability research and exploit development. Some groups reportedly use AI to rapidly analyze security flaws and test proof-of-concept exploits at a scale that would normally require much more time and manpower.
While AI can clearly be useful for productivity and software development, reports like this show how quickly the technology can also be turned into a cybersecurity threat. And according to Google’s researchers, this may only be the beginning.
Written by
