Google says a cybercrime group used an AI tool to help develop a working zero-day exploit that could bypass two-factor authentication — a documented example of AI moving deeper into real-world offensive cyber development.
WHAT’S HAPPENING
Google Threat Intelligence Group says someone inside cybercrime group TeamPCP used an AI tool to help develop a previously unknown exploit targeting widely used login software.
The exploit was designed to bypass two-factor authentication.
Google obtained a copy of the code, tested it and found that it worked after minor modifications. The company then warned the software developer, which patched the vulnerability. WIRED
WHY IT MATTERS
Cybercriminals have been using AI for phishing, coding and reconnaissance for some time.
This crosses a different line.
Google says this is the first cybercrime case it has identified involving a zero-day exploit it believes was developed with AI assistance. austinlarsen.me
That means AI was not simply helping someone understand an existing vulnerability.
It was part of the process used to develop an attack against a flaw that defenders did not yet know existed.
WHO BENEFITS
Security teams gain clearer evidence of how attackers are beginning to use AI inside real offensive workflows.
Google was able to analyze the exploit, confirm the vulnerability and notify the affected developer before broader exploitation could occur.
That gives defenders another reason to use AI for vulnerability discovery, code review and rapid patching.
WHO LOSES
Software vendors and security teams face a faster threat cycle if AI allows attackers to search for and develop exploits more efficiently.
Traditional defenses also become more difficult when attackers can combine stolen credentials, supply-chain compromises and AI-assisted exploit development.
TeamPCP itself allegedly compromised more than 1,000 organizations and exposed more than 500,000 credentials through a broader software supply-chain campaign. Two alleged participants were arrested in Australia, and one was later indicted in the United States. Department of Justice
WHAT HAPPENS NEXT
The key question is whether AI-assisted zero-day development remains rare or becomes a repeatable capability.
Google says adversaries are already moving beyond simple chatbot use and embedding AI directly into attack workflows. austinlarsen.me
The cybersecurity race is changing:
AI is not just helping defenders find vulnerabilities.
Attackers are beginning to use it to find them too.