APREX · EST. 2026
Google's major developer conference placed artificial intelligence at the center of almost everything the company does. From search to email, from photos to video – AI is now woven in as a foundational layer across Google's product universe. But the more data the AI has access to, the more pressing the question becomes: what exactly happens to your information?
Gemini Wants to Know You – The Whole You
Google's large language model Gemini is at the core of the new strategy. Via a feature called “Personal Intelligence,” Gemini can now pull data from Gmail, Google Photos, search history, and YouTube activity to provide more tailored answers. Google emphasizes that the feature is turned off by default, but according to Stratechery, the company is actively encouraging users to activate it.
This creates what privacy experts call “data bleeding.” Miranda Bogen, Director of the AI Governance Lab at the Center for Democracy and Technology, points out that when personal content flows between services and into AI systems' conversation history, it becomes increasingly difficult to keep track of what is happening – and even harder to set limits on what should not happen.
When personal content is integrated across products, it's harder and harder to keep track of what's actually going on – let alone draw lines.
Opt-in in Theory, Opt-out in Practice?
Google assures that the company's core models are not trained on private Workspace data without explicit permission. At the same time, Gemini can be trained on user interactions – including responses that may contain summaries from emails and files. How effectively personal information is filtered out in this process is not publicly documented.
Users have reported that Gemini Spark feels “always on,” and that the AI reads private emails and calendar entries without consent feeling genuine. Critics characterize Google's consent practices as “opt-in initially, and 'no' means 'maybe later'.”
AI Overviews in Search – And Questions About Source Quality
In Google Search, “AI Overviews” – AI-generated summaries – are now a permanent part of the results display. Critics fear that this weakens users' ability to assess where the information actually originates, as fewer links to original sources are shown compared to traditional search.
It is also not unproblematic that AI summaries now directly cite social media like Reddit. The question of the reliability and contextualization of such sources has been raised by several media observers. Previous errors – such as when Google's AI recommended glue as a pizza topping – still cast a shadow over the system's reputation.
DeepMind and Google's Business Goals – The Same Course?
Another underlying tension highlighted by Stratechery is the relationship between DeepMind and Google's commercial ambitions. The question of whether the research division operates in sync with – or independently of – Google's core business has not been answered. It is a question that will become increasingly relevant as AI systems take up more space in products that directly affect billions of users.
Technology is Ready – Regulations Lag Behind
The rapid integration of AI across Google's product portfolio puts existing privacy legislation under pressure. Experts point out that concepts such as “processing” and “transparency” must be redefined to keep pace with what is now happening in practice. Until that happens, much of the responsibility rests with users themselves – who in many cases lack both the tools and the information needed to make truly informed choices.