Social media giant Meta is postponing the launch of its new artificial intelligence (AI) software in Europe following concerns from data protection agencies.
The Irish data protection authority responsible for Meta in the EU demanded the company not train AI models using publicly available posts on Facebook and Instagram for the time being.
Without training using local content, Meta cannot offer users in Europe a first-class experience, the company said in a blog post on Friday.
At the same time, Meta said it is convinced that its approach complies with European laws and regulations.
Meta is not alone in training artificial intelligence - "and we're more transparent than many of our industry counterparts," the company said.
The step follows criticism that Meta only provided an option that people object to the use of their data but did not provide for explicit user consent.
Data protection activists from the organisation noyb then filed complaints in 11 countries last week, and criticized what they called "hidden and misleading" opt-out procedure.
Meta originally claimed that there was a "legitimate interest" in training the AI models and that only publicly available content from adult users should be used for this purpose.
On Friday, the Irish data protection authority welcomed Meta's decision to postpone its plans in Europe.
With the AI software Meta AI, the company is competing against other programmes such as chatbot ChatGPT.
Meta AI is designed to create texts and images and answer users' questions. The Facebook Group has repeatedly emphasised in recent months that artificial intelligence will significantly change platforms.
At the same time, Meta said it intends to continue bringing AI functions to Europe.