Traditionally, the data that technological ones collect from users have served to to do business segmenting the advertising they show on their platforms. However, AI is changing many things and one of them is that these data now have a new value for companies that are dedicated to it, which is that of train their models. Meta has already recognized that he has used the public content of users in Facebook and Instagram, Posted since 2007to that end, but now it has taken a step that takes things much further. The app of Facebook has begun to request users at USA and Canada that Give access to the Meta AI Not only to the content that users have uploaded to Facebook servers, but to the personal images they keep on their devices. Not to get too much attention, the platform has baptized this option as cloud processing.
The notice appears when the user wants to publish a History On Facebook. The app then shows an emerging message asking if the user wants to activate cloud processing for ‘get creative ideas made for you from your camera’s reel ‘.
Reply affirmatively allows Facebook ‘Select content of your camera reel and upload it to our cloud regularly’ to generate ‘ideas such as collages, summaries, style changes through AI or topics such as birthdays or graduations’.
When activating this function, the message continues, users They accept the terms of goal AI in North Americathat allow your AI to analyze ‘Content and facial features’ of those unpublished photos, as well as the date they were taken and the presence of other people or objects in them. In addition, users They grant the right to ‘conserve and use’ that personal information.
Goal tries to dispel the doubts that users can have on this use of their personal images indicating that ‘They will not be used for advertisement segmentation ‘ And, also, what They will be reviewed ‘for security and integrity reasons’.
María CubetaMeta spokesman, has pointed out to TechCrunch that the function is a test and that ‘we are exploring ways to facilitate the shared use of Facebook content through content suggestions ready to share and selected from the reel of the person’s camera’. Also that ‘these suggestions are only optional and only the user is shown, unless deciding In this test‘. We’ll see when it stops being a test.
To the Verge questions about this use of the private images of the users, a goal is reaffirmed that currently (again, the nuance) is not training its AI models with those photos, but He did not respond to whether he could do it in the future or what rights will you have on the images of the camera reel.
This option can be deactivated from Facebook settings, in Preferences. In the section Reel shared suggestionswe must touch the slider corresponding to the cloud processing to leave without effect the permission granted previously.
Will we see it in Europe? It is unlikelysince the laws are more restrictive with what technological ones can do with user data. But it makes it clear that to the risks of artificial intelligence we must add that of how it puts at risk the privacy of users.