Facebook, the social network platform belonging to Meta, asks users to download photos of their phones to suggest collages, summons and other ideas using artificial intelligence (AI), including those that have not been directly downloaded from the service.
According to Techcrunch, which reported for the first time Functionality, users receive a new contextual message requiring permission to “allow the treatment of the cloud” when they try to create a new story on Facebook.
“To create ideas for you, we will select the media from your camera bread and download it on our cloud on a continuous basis, depending on information such as time, location or themes,” notes the company in the pop-up. “You alone can see suggestions. Your media will not be used for targeting the ads. We will check it for safety and integrity purposes.”
If users agree to process their photos on the cloud, Meta also indicates that they accept her IA termswhich allows him to analyze their media and facial features.
On a help page, Meta said “This feature is not yet available for everyone”, which is limited to users of the United States and Canada. He also underlined in Techcrunch that these AI suggestions are OPT and can be disabled at any time.
Development is yet another example of how companies run to integrate AI functionality into their products, often at the cost of user confidentiality.
Meta says that her new AI feature will not be used for targeted announcements, but experts still have concerns. When people download photos or personal videos – even if they accept – it is not clear how long this data is kept or which can see them. Since the treatment occurs in the cloud, there are risks, especially with things such as facial recognition and hidden details such as time or location.
Even if it is not used for advertisements, this type of data could always be found in training data sets or be used to create user profiles. It’s a bit like putting your photo album back to an algorithm that quietly learns your habits, preferences and models over time.
Last month, Meta began To train your AI models Received approval of the Irish Data Protection Commission (DPC). Business suspended The use of generative AI tools in Brazil in July 2024 in response to confidentiality raised by the government.
The social media giant has also added AI features to WhatsApp, the most recent being the possibility of Summarize unrelated messages In cats using a confidentiality -oriented approach, he calls private treatment.
This change is part of a larger trend in the generative AI, where technological companies mix convenience with follow -up. Features such as collages made in automatic or suggestions of intelligent stories may seem useful, but they count on the AI that looks at how you use your devices – not just the application. This is why confidentiality parameters, clear consent and limitation of data collection are more important than ever.
The Facebook AI function also came from one of the German data protection surveillance dogs called Apple and Google to delete Deepseek applications from their respective application stores due to illegal user data transfers in China, after similar concern raised by several countries at the start of the year.
“The service processes extensive personal data of users, including all text inputs, chat stories and downloaded files as well as information on the location, the devices used and the networks”, according to a statement published by the Berlin Commissioner for Data Protection and Freedom of Information. “The service transmits the personal data collected from users to Chinese processors and stores them on servers in China.”
These transfers violate the general data protection regulations (GDPR) of the European Union, given the lack of guarantees that the data of German users in China are protected at a level equivalent to the block.
Earlier this week, Reuters reported That the Chinese AI company helps military and intelligence operations in the country and that it shares user information with Beijing, citing an anonymous official from the US State Department.
A few weeks ago, Openai too landed A 200 million dollars with the US Defense Ministry (DOD) to “develop prototype capacities of the border AI to meet critical national security challenges in the fields of war and business”.
Business said It will help the Pentagon to “identify and prototype the way in which the AI border can transform its administrative operations, to improve how members of the service and their families obtain health care, to rationalize the way in which they examine the data of the program and the acquisition, in support of the proactive cyber-defense”.