Finish announced On Monday, he trained his AI models in the field of public content, such as posts and comments on Facebook and Instagram, in the EU after prior stopping plans in response to regulatory pressure due to data privacy problems. He said that the company would start training its artificial intelligence on user content in the EU. Meta AI users interactions will also be used to train its models.
The announcement appears after a limited version of the meta AI launched in the EU last month, after the debut on the US markets and other global markets.
Although the finish has been training its artificial intelligence for years on the content of users generated by users in the US, it faces the EU resistance due to strict provisions on the privacy of the block, especially the general regulation on data protection (GDPR), which requires a clear legal basis for the processing of personal data to train AI models.
Meta said in June 2024 that he would stop plans to start training his AI systems using user data in the EU and Great Britain after the Irish Data Protection Commission (DPC). DPC regulates the finish in the EU and acted on behalf of several data protection bodies throughout the block. In September 2024, Meta said that he again launched efforts to train AI systems using public posts from the British user base.
Quick to this day, Meta announced that he would do it with public posts with the EU Base.
“Last year, we delayed the training of our large language models using public content, while the regulatory authorities explained legal requirements,” said the finish line in their blog post. “We accept the opinion issued by EDPB, which confirmed that our original approach has met our legal obligations.
Starting from this week, users in the EU will start receiving notifications in the application and e-mail to explain that the finish will start using public data and interaction with meta AI to train their models. These notifications will contain a link to the form that will allow users to opt out of the data used. Meta claims that it honors all forms of opposition he has already received, as well as newly sent.
Meta notes that it does not use private messages or public data from users under 18 in the EU, to train his models.
“We think that we are obliged to build artificial intelligence, which is not only available to Europeans, but is built to them,” says Meta. “That is why it is so important that our generative AI models are trained in various data so that they can understand the amazing and diverse nuances and complexities that make up the European communities. All this, from dialect and colloquialization, to knowledge about hyper-local and different ways in which different countries use humor and sarcasm.”
Meta claims that it follows the example set by companies such as Google and Opeli, which have already used data from European users to train their AI models.
Meanwhile, DPC does not go completely from control how large language models develops their AI services. Last week, the regulatory authority announced that he had studied GROK XAI training.