Google Chrome Criticized Over Hidden Gemini Nano AI Downloads

Reports that Google Chrome has been quietly downloading a large AI model called Gemini Nano onto some users’ devices have triggered a growing backlash among privacy researchers, cybersecurity commentators, and frustrated users online.

The controversy centers less on the AI model itself and more on how it was reportedly installed — with critics arguing many users were never clearly informed that multi-gigabyte files could suddenly appear on their computers in the background. Researchers say the downloaded file, often labeled 'weights.bin,' can consume roughly 4GB of storage space.

Security researcher Alexander Hanff — widely known online as 'That Privacy Guy' — claimed Chrome was downloading the model automatically without explicitly asking for permission first. The issue gained momentum after users began discovering large hidden files buried inside Chrome data folders connected to Google Gemini Nano.

What might have initially looked like a technical detail quickly turned into a wider debate about transparency, consent, and how aggressively major tech companies are integrating AI into everyday software.

What Gemini Nano really does

Gemini Nano is part of Google's plan to put AI directly into consumer devices instead of relying on cloud servers. Unlike bigger AI systems that process requests remotely, Gemini Nano is meant to run on a user's computer or smartphone. Google has been adding the model to Chrome so that it can do things like help with writing, find scams, and use AI to improve autofill and other browser tools.

Supporters say that local processing can improve privacy because sensitive information may stay on the device. However, critics argue that many users may enable these features casually without realizing that doing so triggers the installation of several gigabytes of AI model files behind the scenes.

Privacy concerns quickly escalated

Privacy advocates and cybersecurity researchers responded sharply once reports of the downloads spread across forums. Some critics asked if downloading large AI models without clear permission could break privacy laws in Europe where the GDPR and ePrivacy laws are in place. The criticism focused on lack of explicit consent, hidden storage usage, and increased bandwidth consumption.

Others talked about the bigger effects on the environment. If AI models are being downloaded and maintained across hundreds of millions of devices globally, the combined impact on electricity usage and computing resources could become enormous over time.

Not everyone agrees the situation is malicious

At the same time, some technology analysts argued the strongest criticism may be overstating the situation. Several reports pushed back against descriptions calling Chrome’s behavior 'spyware'. Supporters of Google’s approach noted that on-device AI models are technically necessary for local AI features to function.

A preview of a much bigger industry debate

The controversy arrives during a major industry-wide shift toward embedding AI directly into mainstream software. Google is aggressively integrating Gemini across its ecosystem, while competitors like Microsoft and Apple are pursuing similar strategies. This transition is creating new questions about transparency, consent, and user control.