Meta will soon use your public posts to train its AI. Can you prevent this?

This month, new messages appeared on European users of Instagram and Facebook mobile apps informing them that from June 26, their public posts may be used to train artificial intelligence (AI).

According to Instagram's help center, the information used may include posts, photos, captions and messages sent to the AI, but it will not include the content of private messages.

Parent company Meta said in a statement last week that it would begin to inform people in the UK and EU about “how public information shared in Meta products and services is used to develop and improve Meta's AI, within the bounds of their respective privacy laws.”

Is this GDPR compliant?

According to the General Data Protection Regulation (GDPR), companies and organisations can process personal data if they have a “legitimate interest” to do so.

Mehta noted the legal basis for this, adding that the data would be used to “develop and improve AI.”

You can opt out of this use of your data by filling out the form. Facebook – Currently unavailable – or Instagram.

“We will consider your complaint in accordance with the relevant data protection laws and, if your request is respected, it will apply going forward,” the filing said.

Although you may opt out, your data may still be used by Meta if you appear in an image someone shares or if you are mentioned in another user's post or caption.

The Irish Data Protection Commission (DPC) told EuronewsNext that “Meta has delayed its launch for a number of reasons.” [of] Responded to inquiries from DPC.

The DPC said Mehta notified users four weeks before the initial training.

The technology company also informed the DPC that “only personal data that EU users publicly make available on Instagram and Facebook (posts, not comments) during training will be used, and will not include personal data from the accounts of users under the age of 18.”

Driving AI

In a press release last month, Meta said it was “aggressively” investing to support AI research and product development efforts.

The company has its own large-scale language model (LLM) called Llama, the latest version of which (Llama 3) was released in April and is used to power its assistant Meta AI, which is not yet available in Europe.

According to Reuters, Mehta was already using Instagram and Facebook posts from his public profile to train his assistants.

However, this information was not included in the dataset on which Llama 2 was based.

The company is also working on the infrastructure needed for AI workloads, unveiling custom-made chips last month.