Decide carefully as the new upgrade has confirmed it.
Republished on June 29 with a new IA confidentiality classification and Google’s ranking.
The 2 billion Google Gmail users face a critical decision because the company improves the main global messaging platform to use AI more. This means access to AI based on Google cloud all Your content – as personal and sensitive.
I warned before Gmail users must understand the risks before using all these new updates by default. It is not an accident or a surprise that the upgrades of the Gmail AI are in conflict with its almost end -to -end encryption upgrade. One is secure, the other is not.
The problem with Smart Ai Search and Smart Ai Resplus was confidentiality – pure and simple. This week, all Android users were struck by similar warnings, because a badly written e -mail of Google seemed to suggest that AI would access sensitive applications – including messages and WhatsApp – even if the gemini activity is deactivated.
“Gemini will soon be able to help you use the phone, messages, WhatsApp and public services on your phone, whether your Gemini Apps activity is activated or deactivated,” the e-mail told users. As Gizmodo Report: “Google to Gemini users: we will watch your texts, whether you like it or not.” And of course, everyone has read this email in the same way.
Google now told me that it was misleading – the update “is good for users”. Android phones “will use Gemini to perform daily tasks on their mobile devices such as sending messages, launching telephone calls and defining timer while the activity of Gemini applications is deactivated.”
What this really means is that the deactivation of the Gemini activity does not prevent it from working on phones, which would have happened before. But “with the activity of gemini applications deactivated, the cats of Gemini are not examined or used to improve our AI models.”
Before, if you wanted to use gemini with messaging applications, you should keep the gemini activity on which would be means these interactions saved. It is a nightmare of confidentiality and that is what Google has in fact repaired. This choice East Good news.
This welcome clarity reflects user concerns. But this is what is still lacking in the debate on Gmail AI, which was more an offer of taking or leaving the offer. I would also like to see the same choice of transparent and easy confidentiality for Gmail, before it goes much further.
A note of prudence. Any use of gemini on Android always records interactions for 72 hours in your account, even if Activity is extinguished. Keep this in mind.
And confidentiality warnings have not disappeared. By Android titles“Google’s GEMINI AI will soon be able to access applications such as the phone and messages, even if” the activity of the Gemini applications “is deactivated, from July 7, 2025. Although Google specifies that the deactivation of the activity always prevents data from being used for the training of TIA data, the change in concern of user confidentiality, balancing confidence with confidence in confidence in the management of data Google.”
This new report warns that “more in -depth integration inevitably puts legitimate concerns in the foreground for many users. The prospect of an AI having access, even if it is temporary, to very personal data in call logs, private messages and WhatsApp cats immediately increase red flags concerning individual confidentiality and overall security of data. “
Gen Ai and LLM Data Confidential Ranking 2025
It was always going to happen – it was clear from the first days of integration of Gemini into emails and messages and other applications, just as it is clear with the news that WhatsApp must do the same With messages from its own users.
As Futurism The said: “WhatsApp now offers summaries of AI of text sons for those who are too lazy to read their messages themselves … because what says” I care “to break a summary generated by the AI of the cat of the family group?
A balance must be found and the risk is that users do not yet know enough about the complex network of privacy policies to make these decisions. And this is particularly true with regard to young users, who have not yet developed confidentiality awareness.
But now as Techradar Prevents: “Google deploys Gemini for schools and students in the midst of concerns for the future of education around the world”, which means “Gemini arrive in a school near you, whether you wanted it or not.”
With Gemini now promising “to help everything, from the planning of lessons to the comments in real time, [its] The expansion also raises difficult questions about the long-term role of AI in education and how it can reshape learning itself. The risks are high.
In his new report “Gen Ai and LLM Data Privacy Ranking”, Incogni Provides some tips on how Gemini really accumulates. “Although these tools can improve productivity,” he warns, “most users are not aware of the complex confidentiality challenges behind the scenes”. And this is particularly true with regard to Google, Microsoft and Apple Deployment of AI almost by default in the operating systems used by billions.
“As these sophisticated models are becoming more and more integrated into daily work flows”, explains Incogni, “from content creation to code generation – the unauthorized data sharing potential, abusive use and exposure to personal data has increased more quickly than childcare dogs or confidentiality assessments can follow.”
Unsurprisingly, given the landscape of harvesting and monitoring reports of wider data, Meta ranks last in the list with Google not much better, second in second place above the Microsoft co-pilot. Chatgpt does much better, which will play in the new story that users prefer this LLM to Big Tech offers.
“Maintaining awareness of the evolution of confidentiality risks and data processing practices has simply become impractical for the average user,” according to researchers. “The confidentiality assessments of LLMS and Gen AI processes acquire their training data and the sensitive information exposed thanks to the current interactions of users with them.”
Google ranks much better with regard to the data used to form models and for global transparency. And this goes to the new deployment of education.
“Confidentiality and security have been highlighted by Google”, ” Techradar Said: “Gemini for Education follows the same data protection as the work space for education. In a word, this means that students’ data is not used to form AI models or examined by humans.”
