EU mandates labeling for AI-generated content, fines for non-compliance

The European Union has introduced new regulations mandating the clear labeling of content created by artificial intelligence. The primary goal is to ensure that Europeans can instantly distinguish between authentic and fabricated online material.

“

Generative AI creates the conditions for creating disinformation tailored to specific audiences on an unprecedented scale and spreading it extremely quickly.

Under the EU's AI Act, companies must now clearly inform users when they are interacting with AI, such as chatbots. Furthermore, any text or images generated by AI must be accompanied by appropriate labels, which can include digital watermarks or other identifiers. This measure aims to combat the spread of disinformation and maintain public trust in online content.

Companies that fail to adhere to these new requirements risk substantial penalties. Existing AI systems have until December 2 to align with the regulations. The EU has expressed particular concern over "deepfakes" – realistic-looking but entirely fabricated texts, images, or audio created using AI.

“

It is becoming increasingly difficult for all of us to distinguish between what is real and what is artificial, (therefore the new EU rules aim to) preserve the ability of citizens to trust what they see, hear, and read.

While the rules generally apply to content created for professional purposes, they do not affect individuals using AI for personal use. However, text intended for public information on matters of public interest must also be labeled if generated by AI and not human-reviewed. Exceptions are made for "artistic, creative, satirical, or fictional" works.

“

We heard that this will be very, very difficult to implement. However, I think we often hear that when talking about compliance requirements. Nevertheless, the world does not stop, and we manage to solve these problems.