European Parliament committees have approved the first set of technical rules to put the EU's AI Act into practice, moving the bloc closer to enforcing new requirements on artificial intelligence companies. The measures are among the earliest steps in turning the landmark law into day-to-day rules for developers and platform operators.

According to Reuters, the proposed standards would require large language model providers to be more transparent about the sources used to train their systems. They would also need to build stronger safeguards aimed at moderating harmful content in real time, a provision that has drawn resistance from parts of the tech industry.

Companies affected by the rules have argued that the obligations could be difficult to implement at scale and may create operational burdens. Supporters of the framework say the requirements are necessary to reduce risks tied to misinformation, abuse, and unsafe outputs from rapidly deployed AI tools.

The committee vote does not complete the regulatory process, but it marks a significant step in the EU's effort to set enforceable standards for artificial intelligence. As the law moves toward implementation, the debate is likely to intensify over how to balance innovation, transparency, and public safety.