Hello everyone, that AI model of four gigabytes that you didn’t request? Guess what! It has completed its download onto your computer. The uninvited software that consumes data? It’s just another way Google is monitoring you, ensuring that—oh. You prefer not to have an AI imposed on your system without your consent, and you believe Google should know better? Well…

Security expert Alexander Hanff, also recognized as “That Privacy Guy,” recently disclosed that Google’s popular Chrome web browser is covertly saving an on-device AI model without obtaining explicit user approval. As per Hanff, Chrome discreetly caches approximately 4GB of AI model data on the devices of Chrome users (although many users have had difficulty locating the file). The said model is stored in a directory labeled OptGuideOnDeviceModel, with the main component being a file named weights.bin. Hanff alleges that these weights are for Gemini Nano, which is Google’s less resource-intensive on-device version of its extensive language model.

This scenario serves as yet another instance where nothing comes free when it comes to the internet or the intrusive practices of mega corporations. At present, there appears to be no straightforward method to eliminate the local AI from the hosting system without uninstalling Google’s Chrome browser.

Consider this a signal that it might be a suitable time to transition to a different browser. For instance, Firefox can be freely downloaded. Similar to Chrome, it can be tailored extensively using add-ons and themes, offering numerous enhancements to safeguard your privacy. Notably, in relation to this story, Firefox’s parent company, Mozilla, empowers users with the option to decide whether or not the browser utilizes AI functionality.