Google Search Engine Exposes Bard AI User Privacy


According to The New York Times, the accuracy of Google's AI Overview feature is about 90%. With Google's annual search volume exceeding 5 trillion searches, this means that millions of incorrect answers may be generated every hour, and nearly a million pieces of incorrect information per minute. An assessment by startup company Oumi showed that the accuracy of Google's Gemini model increased from 85% in October last year to 91% in February this year.
Google plans a 933 MW natural gas plant in Texas to power AI data centers, raising concerns over its zero-carbon pledge.....
Google has released the new open source AI model Gemma4, which adopts the Apache 2.0 license, replacing previous restrictive agreements. This allows developers to freely use, modify, and distribute the model, facilitating commercial applications. The model achieves dual upgrades in technical architecture, improving performance and ecosystem compatibility.
Perplexity AI faces a U.S. class-action lawsuit for allegedly sharing sensitive user data with third parties like Meta and Google without consent, potentially exposing private financial and tax information even in incognito mode. The company states it has not yet received formal legal documents.....
Google unveils the lightweight video generation model Veo 3.1 Lite, significantly reducing computing costs through architectural optimization, aiming to promote the popularity of high-quality video generation. The model targets lightweight market products, offering high cost-effectiveness and a video generation cost as low as $0.05 per second for 720P videos, meeting the needs of different developers.