JAKARTA - Meta Platform Inc has confirmed it will delay the company's major language model (LLM) training using European user data, as it is subject to applicable regulations there.

"We are disappointed with the request from the Irish Data Protection Commission (DPC), our main regulator, on behalf of the European DPA, to postpone our large language model (LLM) training using public content shared by adults on Facebook and Instagram," the company wrote in its latest announcement.

According to Meta, this is a setback for innovation in Europe, competition in AI development, and further delay in providing AI benefits to people in Europe.

Meta also believes that it is not only those who use user data to train their language models. Nevertheless, Meta remains committed to complying with applicable laws and regulations. That is, Meta will not be releasing Meta AI in Europe for now.

Meta will also continue to collaborate with the DPC so that Europeans can get access to good AI innovations.

The delay will also allow Meta to handle specific requests from the Information Commissioner's Office (ICO), a company regulator in the UK, before starting training.

Earlier, on June 10, Meta announced it was building artificial intelligence (AI) technology specifically for Europe that reflected linguistic, geographical, and cultural references.

At that time, Meta admitted that it would train their big language model using user-selected content in the European Union to be shared publicly on its platforms such as Instagram and Facebook.


The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)