:

GOOGLE CHROME INSTALLS 4GB AI MODEL WITHOUT USER CONSENT

AI DESK2 MIN READ
TUE, MAY 5, 2026

■ AI-SUMMARIZED FROM 1 SOURCE BELOW

Google Chrome has been automatically installing a 4 GB AI model on user devices without explicit permission. The installation occurs silently in the background, raising privacy and storage concerns.

Google Chrome is deploying an undisclosed AI model to user devices without active consent, according to reports from privacy researchers. The model, approximately 4 GB in size, installs automatically during browser updates. What's happening The AI model appears to be integrated into Chrome's latest versions. Users report finding the large installation on their systems without initiating any download or accepting specific terms for the feature. The silent deployment bypasses traditional consent mechanisms. Storage and privacy implications The 4 GB footprint represents a significant addition to Chrome's resource demands, particularly for users with limited storage or slower connections. The automatic installation raises questions about data collection, processing, and how the model operates on local devices. Privacy advocates note the lack of transparency around the model's purpose, data handling, and operational scope. Users receive no clear notification of what the AI does or how it affects their system performance. User response Discussion on Hacker News generated substantial engagement, with 177 comments exploring removal methods and broader implications. Users expressed frustration over the inability to opt-out before installation. Broader context The installation aligns with Google's push to integrate AI capabilities into Chrome. The company has previously announced AI features for the browser, though the silent deployment of a 4 GB model represents a more aggressive implementation than typical feature rollouts. Chrome users seeking more information can review detailed analysis and discussion at privacy-focused tech outlets and developer communities. The incident underscores ongoing tensions between feature deployment at scale and user transparency expectations. Google has not issued an official public statement regarding the automatic installation or provided opt-out mechanisms at the time of reporting.

■ SOURCES

Hacker News

■ SUMMARY WRITTEN BY AI FROM THE LINKS ABOVE

■ MORE FROM THE SECURITY DESK

Limited access to advanced defensive AI systems is leaving companies, central banks, and nations exposed to increasingly sophisticated AI-powered cyberattacks. The disparity in cybersecurity capabilities threatens global digital infrastructure.

1H AGOAI Desk

A Latvian national extradited to the U.S. has been sentenced to 8.5 years in prison for negotiating ransoms as a "cold case" operator for the Russian Karakurt extortion gang.

1H AGOIndustry Desk

APT37, the North Korean hacker group also known as ScarCruft, has distributed an Android version of the BirdCall backdoor through a compromised video game platform in a supply-chain attack.

2H AGOSecurity Desk

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has issued a warning about a severe vulnerability called CopyFail affecting major Linux versions. The bug is currently being exploited in active hacking campaigns targeting servers and datacenters.

4H AGOAI Desk

■ SUBSCRIBE TO THE DAILY BRIEF

ONE EMAIL, 5 STORIES, 06:00 UTC. UNSUBSCRIBE ANYTIME.