Calif Research's AI-built WeChat worm spread with zero user interaction
The case shows AI materially compressing the time and team size needed to build cross-platform zero-click malware. The disclosure also put a private firm in the position of managing a sensitive cross-border vulnerability that US government policy complicated.
Calif Research released a demo of WeWorm, a self-propagating worm that infects a device by placing a WeChat call, requiring no answer or interaction from the victim. Once on a device, it reads and sends messages, places calls, and spreads to all saved contacts. Calif described it as the first known worm to spread across both iOS and Android without any user click or tap. Researchers estimated it could have hijacked millions of WeChat accounts within hours. Chained with other bugs, Calif said it could fully compromise an infected device. Tencent patched the vulnerability after Calif reported it; WeChat has 1.4 billion monthly users and Tencent says no users were affected.
Calif used AI to find the underlying bug and produce a working remote code execution exploit in about two days, with the full worm complete in roughly one additional week. Calif said: "A worm at this scale used to be the kind of thing that took a larger team months. AI can already do most of the work here. Our team provided the judgment about what to target and how to test it safely."
Calif reported the vulnerability to Tencent directly despite the Pentagon's ban on working with Tencent due to alleged Chinese military ties. A ChinaTalk analysis argued that private companies are better positioned than governments to manage cross-border AI and cyber threats, noting that "cyber vulnerabilities don't map neatly onto state borders." That analysis described US-China trust on AI safety as nearly nonexistent, with the two sides not agreeing on a common definition of AI safety, and said a technically-driven private-sector approach could build a stronger foundation for bilateral agreements than ideologically-driven government efforts.
A September 11 report on Calif Research's zero-click WeChat worm, WeWorm, added that the tool could have hijacked millions of WeChat accounts within hours, a scale claim absent from earlier coverage of the disclosure.
Tencent patched the underlying vulnerability after Calif's alert and said no users were affected.
Calif Research published a demo on September 10 of WeWorm, a zero-click worm that spreads through WeChat calls on iOS and Android with no user interaction required.
AI found the bug and wrote the first remote code execution exploit in about two days, completing the full worm in one additional week, a pace Calif Research said previously required a larger team and months of work. Tencent patched the flaw through private-sector contact Calif initiated despite the Pentagon's ban on working with Tencent.
September 10 reporting added that WeWorm, after infecting a device through a WeChat call, reads and sends messages, makes calls, and spreads to all saved contacts, and could fully compromise the phone if chained with other bugs.
Tencent said no users were affected among WeChat's 1.4 billion monthly users. A ChinaTalk analysis argued Calif's direct contact with Tencent despite the Pentagon ban shows private companies may handle cross-border AI and cyber threats better than governments, with US-China trust on AI safety nearly nonexistent.
I send one email each morning with the stories that moved. If you would rather just read here, that works too.
Subscribe free