Post · bonfire.cafe

“Google Chrome is reaching into users' machines and writing a 4 GB on-device AI model file to disk without asking. The file is named weights.bin. It lives in OptGuideOnDeviceModel. It is the weights for Gemini Nano, #Google's on-device #LLM. #Chrome did not ask. Chrome does not surface it. If the user deletes it, Chrome re-downloads it.”

#privacy #sustainability #consent #values

https://www.thatprivacyguy.com/blog/chrome-silent-nano-install/

That Privacy Guy!

Google Chrome silently installs a 4 GB AI model on your device without consent. At a billion-device scale the climate costs are insane. — That Privacy Guy!

Google Chrome is downloading a 4 GB Gemini Nano model onto users' machines without consent, with no opt-in, no opt-out short of enterprise tooling, and an automatic re-download every time the user deletes it. The pattern is identical to the Anthropic Claude Desktop case I wrote about last month, but the scale is between two and three orders of magnitude larger. This article does the legal analysis and, for the first time, the environmental analysis. The numbers are not small.

#values #consent #sustainability #chrome #google #llm #privacy

Fenix

@fenixmaster@mastodon.social · 6 minutes ago

@wlaatje WTF, fortunately I don use Chrome, but Brave, albeit Chrome based.

Crisper

@Crisper@techhub.social · 2 hours ago

@wlaatje its easy, dont use chrome.

Sibrosan

@sibrosan@mastodon.social · 2 hours ago

@wlaatje het is niet dat ik nóg een reden nodig had om Chrome niet te installeren...

Erik Jonker

@ErikJonker@mastodon.social · 3 hours ago

@wlaatje ...using Google all my life and embedded in it's ecosystem... (i am planning to get out), so for me this is not really surprising. The most worrying aspect for me is if it does not actually contribute to my Chrome experience/functionality (part of your argument). Moving models to "the edge" in itself is not strange.

Wladimir Mufty

@wlaatje@social.edu.nl · 2 hours ago

@ErikJonker 👌indeed! Apart from the fact that it was unsolicited, there is also the question of whether millions/billions(?) of clients should be equipped with this capability. That may become a future scenario in which energy, can no longer be taken for granted to power (or cool) massive centralized AI data and compute centers 24/7. Local GenAI on laptops also consumes electricity, of course, but it is decentralized and only needs to be used when people actually make use of it, not 24/7.