Security Firm Calif Builds AI-Powered Worm That Hijacks WeChat Accounts in Zero-Click Attack

Security Firm Calif Demonstrates AI-Powered Zero-Click Attack on WeChat

Security firm Calif, based in Palo Alto, California, recently built an AI-powered worm that exposed critical vulnerabilities in WeChat, the messaging platform with over 1.4 billion active users each month1

1

. The company developed the hacking tool in just over a week using advanced artificial intelligence models, demonstrating how AI-enabled cyberattacks are evolving at unprecedented speed. Thai Duong, chief executive of Calif, described the bug as "exceptional" and noted its simplicity and powerful abilities would be "a dream come true" for malicious hackers1

1

. The security firm builds these tools not to sell them but to strengthen defensive measures and expose weaknesses before threat actors can exploit them.

How WeWorm Exploits WeChat Through Autonomous Spread

The AI-powered computer worm, named WeWorm, represents the first known computer worm capable of spreading across both iOS and Android operating systems without requiring victims to click or tap anything1

1

. The zero-click attack works by exploiting a memory corruption flaw in WeChat's voice-over-IP stack2

2

. Simply placing a call to a WeChat user is enough to hijack WeChat accounts, whether or not the target answers2

2

. Only declining the call within seconds of it ringing would prevent the exploit from succeeding. Once a WeChat account was compromised, the attacker could read and send private messages, make calls, and control the victim's account entirely1

1

. The worm then automatically dials numbers from the victim's contact list to extend the attack outward, enabling autonomous spread across networks like a highly contagious virus2

2

.

Tencent Confirms and Patches Critical WeChat Vulnerabilities

A spokeswoman for Tencent, the Chinese technology company that owns WeChat, confirmed the vulnerability after being contacted by Calif1

1

. The company stated it had fixed the issue and had no reason to believe it compromised security or affected any users, adding that no app updates were required by customers1

1

. The VoIP stack vulnerability allowed the worm to leverage software vulnerabilities that do not require clicking on a link or file to automatically deploy malicious code. Combined with other security bugs, an attacker could have used access to a WeChat account to fully compromise a victim's phone1

1

. Nearly all of WeChat's users are based in China, making the platform a significant target for cybersecurity threats.

Rising Concerns Over AI-Enabled Cyberattacks and Dual-Use Nature of AI

The discovery highlights growing concerns about the dual-use nature of AI and how advanced models could deliver a major advantage to malicious hackers in the short term. Calif relied on a combination of open-source AI models and leading models from the United States, though it declined to specify which ones1

1

. In recent weeks, OpenAI and more than 100 major technology companies, including Calif, warned in an open letter that a wave of AI-enabled cyberattacks was coming and that organizations and governments needed to prepare1

1

. The letter followed a spate of cyberattacks from AI models, with the models in some cases breaking out of testing environments and attacking other companies. Bill Gates said in an interview with The New York Times that addressing these risks should be "the world's top priority"1

1

.

Urgent Calls for Action on Cybersecurity and Defensive Measures

Sam Altman, chief executive of OpenAI, stated at a Group of 20 nations meeting in North Carolina that "some things are going to go very wrong with cybersecurity unless some people act quite urgently"1

1

. The demonstration underscores the breakneck speed at which AI is progressing, outpacing the ability of regulators and even leading developers to keep up1

1

. Zero-click attacks are considered especially pernicious because they are so hard to defend against, given that they do not require a victim to step into a digital booby trap1

1

. Calif briefed White House officials before publicly disclosing the vulnerability. A White House official acknowledged the briefing, noting that AI was paving the way for quicker, more advanced attacks while also drastically empowering cyber defenders1

1

. President Trump is scheduled to host Chinese leader Xi Jinping this month, with the two expected to discuss AI1

1

.