Google Chrome silently installs a 4 GB AI model on your device without consent (www.thatprivacyguy.com)

Google Chrome has come under fire for silently installing a 4 GB AI model named "weights.bin" on users' devices without consent. This file contains the weights for Gemini Nano, Google's on-device large language model (LLM), which powers features like "Help me write" and on-device scam detection. The model is added as Chrome's AI capabilities are activated by default, and even if users try to delete the file, Chrome automatically re-downloads it, effectively locking users out of the decision-making process. This situation raises significant concerns for the AI/ML community, particularly regarding user privacy and data protection compliance. The installation is argued to breach multiple European regulations, including the GDPR's principles of lawfulness and fairness. Furthermore, the environmental impact of such mass installations could result in substantial CO2 emissions, estimated between 6,000 and 60,000 tonnes depending on the scale. The incident reflects a broader trend where major tech companies deploy software that crosses vendor trust boundaries without user consent, marking a pivotal challenge for ethical AI practices and regulatory frameworks.