Google Sues Cybercrime Ring That Used Gemini to Build Scam Sites

Key Takeaways

• Google filed its first lawsuit against actors who weaponized Gemini AI for cybercrime
• Outsider Enterprise sent 2.5 million scam texts and created 9,000 fake websites
• The FBI is running a parallel criminal investigation alongside Google's civil suit

Google filed a lawsuit on June 12, 2026, against a cybercrime network called Outsider Enterprise. It marks the first time the company has taken legal action against actors who allegedly abused its Gemini AI to build scam infrastructure.

The China-based group ran what amounts to a phishing-as-a-service operation through Telegram. Customers who lacked technical skills could buy ready-made scam kits, complete with fake websites and text message templates. Gemini did the heavy lifting.

How the Scam Network Operated

According to Google's legal filing, Outsider Enterprise provided step-by-step instructions on how to use Gemini to create convincing fake websites. The AI-generated sites imitated Google, YouTube, and government agencies like New York's E-ZPass toll system.

The group offered nearly 300 scam templates. The text messages typically claimed there was a problem with the recipient's account or an issue with a package delivery. When users clicked the links, they landed on fraudulent websites designed to steal personal data and banking details.

Google has tracked 9,000 fake websites and 1 million URLs connected to the network. The company's filing does not estimate how much money was stolen, but Google says hundreds of people lost funds to the operation.

Carrier Coordination and AI Defenses

Google worked with AT&T, Verizon, and T-Mobile to block many of the malicious text messages. The company also credits its on-device scam detection in Google Messages with reducing successful phishing attempts.

That AI-powered feature blocks 10 billion scam texts every month, according to Google. Some of those blocked messages likely came from Outsider Enterprise, though Google did not provide specific numbers.

FBI Investigation and Enforcement Limits

Beyond the civil lawsuit, Google is assisting the FBI's cybercrime division with a parallel criminal investigation. But there's a catch: no one knows who actually runs Outsider Enterprise.

Even if Google had names, enforcement options are limited when perpetrators operate from China. The lawsuit functions partly as a legal tool to seize domains and disrupt infrastructure, not necessarily to collect damages from identified individuals.

“This legal action marks a new front in the fight against AI-driven fraud, where technology providers take direct legal action against those who abuse their generative models.”

— Industry Analyst, Tech Policy Group

The AI Safety Paradox

Google promotes the security measures baked into Gemini with every new model release. But those safeguards clash with a fundamental requirement: chatbots need to follow instructions and be helpful to users.

That tension creates openings. Scammers found ways to use Gemini for website generation without triggering safety filters. The result was thousands of convincing phishing sites that looked legitimate because they were built with legitimate tools.

Google has sued scammers before, but targeting AI abuse represents new legal territory. The lawsuit may set precedent for how tech companies respond when their own generative AI products become weapons.

What Happens Next

The lawsuit asks the court to shut down Outsider Enterprise's operations and block the domains used in the scam network. Whether that stops the people behind it is another question. Decentralized criminal networks can spin up new infrastructure faster than courts can issue orders.

For now, Google's move signals that AI companies are willing to pursue legal action when their tools get weaponized. It also highlights a problem that will only grow: as generative AI gets better at creating convincing content, the barriers to running sophisticated scams keep dropping.

Frequently Asked Questions

What is Outsider Enterprise?

Outsider Enterprise is a China-based cybercrime network that allegedly operated a phishing-as-a-service business through Telegram, using Google's Gemini AI to create fake websites and scam templates.

How did scammers use Gemini AI for phishing?

The group provided instructions on using Gemini to generate convincing fake websites that imitated Google, YouTube, and government agencies like E-ZPass. These sites were used to steal personal and banking information.

How many people were affected by the Outsider Enterprise scams?

Google tracked 2.5 million scam text messages sent to Android users and 9,000 fake websites. Hundreds of people lost money, though the total amount stolen was not disclosed.

Is anyone being criminally charged?

The FBI is running a parallel criminal investigation, but no specific individuals have been identified. Enforcement is complicated by the fact that the operators are based in China.

Can Google prevent its AI from being used for scams?

Google builds safety measures into Gemini, but these conflict with the need for AI to be helpful and follow user instructions. Scammers found ways to use the tool without triggering filters.

Source: Ars Technica