Meta has announced a delay in launching and training its AI assistant in Europe following a request from the Irish Data Protection Commission (DPC). The regulator asked Meta to postpone training its large language model (LLM) on public content shared by adults on Facebook and Instagram, as stated in a June 10 announcement. Acting as Meta's primary privacy regulator in the EU, the Irish DPC made this request on behalf of European Data Protection Authorities (DPAs).

Due to this delay, Meta stated that it cannot currently introduce Meta AI in Europe, expressing disappointment with the decision and viewing it as a setback for AI competition and innovation within the EU. Despite this, Meta emphasized its confidence in complying with European regulations.

The Irish DPC supported the decision to pause training activities, stating it welcomes Meta's response and will continue to collaborate following extensive engagement. The situation evolved over several months, with Meta informing European DPAs of its activities since March and engaging with the Irish DPC after notifying users of its AI training plans through in-app notifications and emails starting May 22.

Concerns from European privacy organizations and certain DPAs, including complaints from NOYB and others, reportedly prompted the Irish DPC's request. Meta highlighted its consultations with the Irish DPC and adjustments made based on feedback to ensure compliance with AI laws. Despite the delay, Meta plans to align with EU General Data Protection Regulation (GDPR) requirements under the legal basis of "Legitimate Interest," similar to practices employed by Google and OpenAI for AI training with user data.