EU mandates labeling for AI-generated content starting August 2

Starting August 2, the European Union will require companies to label content created using artificial intelligence. This measure, part of the AI Act, aims to increase transparency and help the public differentiate between authentic and AI-generated material. "It is becoming increasingly difficult for all of us to distinguish the real from the artificial," explained an EU Commission official, emphasizing the goal to "preserve the ability of citizens to trust what they see, hear, and read." The AI Act, adopted two years ago, regulates artificial intelligence. Chatbots, virtual avatars, and other AI-driven interactive systems must clearly identify themselves as such. AI system providers will also need to embed digital watermarks or other markers to easily detect AI-generated content. A key measure for the public involves labeling "deepfakes" – texts, images, videos, or sounds that appear authentic but are generated or manipulated by AI. Organizations and individuals publishing such content professionally will be obligated to signal AI involvement. AI-generated texts intended to inform the public on general interest topics, without human editorial oversight, must also be labeled. The EU has provided sample icons (e.g., "AI," "AI-generated," "AI-modified") and a guide for companies. Non-compliance could result in fines. However, existing AI systems have until December 2 to adapt, and exemptions exist for "artistic, satirical, fictional, or creative" works, as well as content edited solely with AI assistance. Some critics argue that Brussels is creating excessive regulatory complexity, potentially leading to the labeling of all content due to AI's widespread use in the European economy.