Is Google using your photos, videos and voice to train AI? Here's how users can take control

Google users may be giving the company’s artificial intelligence systems more data than they realize.The tech giant recently revamped its privacy controls to allow Google to store more user-generated content, including images, videos and audio recordings, to improve its AI models and related services.The change was said to have been communicated to users in an email in June and was touted as a way to give users more control over saved history and personalised recommendations, but the update also came with broader data collection settings that are switched on by default for many users unless they choose to switch them off.The new framework introduces two settings: Search Services History and Personalized Recommendations. These settings determine how activity across Google's services is stored, how long it is retained and how it may be used to personalize experiences and improve Google's products, including AI systems.Importantly, the settings extend beyond Google Search. According to multiple reports, they also apply to products such as Google Maps, Shopping, Flights, Hotels, Translate and Google News. This means information generated while using multiple Google services could potentially be retained and used for AI development purposes.A photograph uploaded through Google Lens to identify an object may be saved. Voice searches conducted using Search Live within the Google app can also be stored. Audio clips generated while practicing pronunciation on Google Translate may likewise be retained.Google has stated that it may use saved media to develop and improve its services, including AI models and safety systems. The company's documentation also notes that saved history can help train generative AI models, sometimes with the involvement of human reviewers.While some data retention is necessary for services to function properly, the report notes that Google acknowledges certain saved media can be retained specifically for AI training purposes.The move is part of a larger trend in the tech industry. Companies that once depended heavily on publicly available web data to train their AI systems are increasingly looking to the content that people create or upload while using everyday products. Meta has taken a similar approach, using user-generated photos, media and content captured through AI-enabled devices to help train AI.For users who want to limit Google's access to this information, the report recommends reviewing the Search Services History and Search Services Personalization settings. Within Search Services History, users can disable the "Save Media" option separately or switch off Search Services History entirely. Google also allows users to choose automatic deletion periods of three, 18 or 36 months for stored activityUsers are also advised to review other privacy controls, including Web & App Activity, Timeline and YouTube History, which continue to influence how Google collects and uses personal data. Google has now separated Search data controls from traditional Web & App Activity settings. As a result, users who previously disabled Web & App Activity may still have Search Services History enabled and may need to manually turn it off to fully opt out.Media reports highlight that user content across services such as Search, Maps and Translate could be used to improve AI systems, underscoring growing concerns around transparency and user awareness as AI becomes increasingly embedded across Google's ecosystem.