Digital Economy

How to deny consent to the use of our data to train Meta Ai

A short guide to preventing our data from being used for AI. The case of Meta

7 June 2024

Aggiungi Il Sole 24 Ore
ai preferiti su Google

4' min read

If we do not like Meta's artificial intelligence using our photos, posts, videos (or other public data) to improve itself, let us beware. We must formally object to this use, which will be triggered by the new Meta privacy policy from 26 June.

It is not easy to find, on Meta's pages, information on how to object, and perhaps not even, for many users, to become aware of this possibility. Or even the fact that Meta intends to use our data. By following this guide, however, we can succeed.

What data does Meta use for AI

First of all, what data does Meta use on Facebook and Instagram? The company writes on its new privacy policy page that it uses "publicly available online data and licensed information". It also uses 'information shared in Meta's Products and Services, e.g. posts, photos and related captions'. At least, it does not use 'the contents of private messages you exchange with family and friends to train our artificial intelligence. For more details on how we use data from Meta's Products and Services, please see our Privacy Policy."

Please note, however. To train our models, we collect public information from the Internet or data licensed from other providers; this information may include personal data. For example, if we collect information from a public blog post, this might include the author's name and contact information. When we receive personal information as part of such public or licensed data that we use to train our models, we do not link this data specifically to any Meta account.

Aren't we on Facebook or Instagram? It still affects us. "Even if you don't use our Products and Services or have an account, we may still process your information to develop and improve Meta's AI. For example, this could happen if you are featured in an image shared on our Products or Services by someone who uses them or if someone mentions information about you in posts or captions they share on our Products and Services."

Why do you use them?

Meta is investing billions on AI and says it really needs our data to improve its artificial intelligence. The aim is twofold, he writes: to improve both the AI that it will integrate into Meta's products (such as chatbots, tools for creating images and advertising campaigns, etc.) and the open source AI of the Llama model.

"We are developing new generative AI experiences and look forward to offering them in Meta Products to more people and companies globally. That's why we are developing Meta AI, our collection of generative AI experiences and features, such as Meta AI and creative AI tools, as well as the models that enable them."

"We are also making models available through an open platform to support researchers, developers and other community members.

"AI at Meta enables people to solve complex problems, increase imagination and create like never before. From real-time chat responses to organising and planning the next holiday, through new ways of expression, AI at Meta helps people improve the activities, experiences and moments of everyday life."

As is well known, generative AI needs a lot of data to improve, in the training phase, and give us better results (textual responses, image creation, video, data analysis...).

OpenAi and Google have 'devoured' internet data for their own models. Meta can do this but also something else that only it can do: use the data published on Instagram and Facebook. The largest social audience in the world. Google also uses YouTube videos to train AI, but only Meta has content created by billions of people in so many formats.

It therefore has a competitive advantage and has every interest in exploiting it.

How to oppose

Meta believes it is entitled to use these data according to the 'legitimate interest' discipline of the European privacy regulation Gdpr.

"We believe that the use of this information is in the legitimate interests of Meta, our users and others. To develop and improve Meta's AI, in the European region and the UK we rely on the principle of legitimate interests to collect and process personal information included in publicly available and licensed sources, as well as information that people share in Meta's Products and Services. In other jurisdictions, where applicable, we rely on an appropriate legal basis for the collection and processing of this data.

But again according to the Gdpr, "you are entitled to rights over the way information is used for Meta's AI". "This includes the right to challenge the information you have shared in Meta's Products and Services used to develop and improve Meta's AI. You can also submit requests regarding your personal information from third parties used for the development and improvement of Meta's AI. Learn more and submit your requests here".

We have to fill in a form that can be found here linked from the privacy policy page.

We have to indicate our country, the mail, why we object, i.e. "explain to us what impact this data processing has on you". Optionally, we can provide other information "that might help us to examine this objection". And then click on Submit to send the form.

Meta does not feel obliged to exclude our data from this processing, but various tests show that it tends to give its consent to users' requests immediately.