Google sues Chinese company for Telegram phishing scams using AI-powered Gemini technology, highlighting cybersecurity threat

Editorial illustration for Google sues Chinese Outsider Enterprise for Gemini-driven phishing on Telegram

Google sues Chinese Outsider Enterprise for...

Google sues Chinese Outsider Enterprise for Gemini-driven phishing on Telegram

By AI Daily Post Edited by Brian Petersen, Editor-in-Chief

June 12, 2026 • 2 min read

Google has filed a lawsuit against a Chinese cyber‑crime outfit called Outsider Enterprise, accusing the group of running a large‑scale phishing operation that leans on Google’s own Gemini generative‑AI model. While the tech is impressive, the filing says the network used Telegram channels to sell “phishing‑as‑a‑service” to people who lack the skills to build fraudulent sites and text campaigns themselves. Here’s the thing: the instructions provided on those channels showed how to harness Gemini to spin up nearly 300 ready‑made scam templates that mimic Google, YouTube and even New York’s E‑ZPass portal.

The result, according to Google, was more than 2.5 million text messages blasted to Android users, with about 55 000 of them sent in a two‑week span last month. In total, Google has traced 9 000 counterfeit websites and a million URLs tied to the operation. The texts typically warned of account problems or missed deliveries, then directed recipients to AI‑crafted pages designed to harvest personal and banking data.

Google says it is working with law‑enforcement agencies and mobile carriers to counter the abuse.

Google worked with AT&T, Verizon, and T-Mobile to block many of these malicious text messages, and Google notes that its on-device scam detection in Google Messages probably helped reduce the number of successful phishing attempts, too.

— Google, Google sues Chinese cybercrime network that used Gemini to automate scams - Ars Technica AI

Why this matters

Google’s lawsuit against the Chinese‑based Outsider Enterprise puts a concrete legal front on the misuse of its Gemini model, reminding us that generative AI can be weaponised as easily as it can be innovated. The filing alleges that the group runs a phishing‑as‑a‑service operation on Telegram, automating fraudulent sites and text campaigns for users lacking technical skill. For developers, this underscores the importance of embedding abuse‑prevention safeguards into any Gemini‑powered product we ship.

Founders should note that reliance on third‑party platforms may expose them to similar liability, especially when law‑enforcement partnerships are invoked. Researchers are left with a clear signal: progress in model capabilities must be matched by rigorous study of how those capabilities can be co‑opted for crime. Yet, the filing offers few details on the scale of the operation or how effective Google’s legal strategy will be; it remains unclear whether this case will deter future abuse or simply shift tactics elsewhere.

As we continue to build with generative AI, vigilance and proactive risk management are essential.