Artificial intelligence is becoming increasingly difficult to distinguish from human-created content. AI can now generate convincing texts, images, audio and video—and that creates enormous opportunities, but also new possibilities for manipulation.
The European Union is responding with stronger transparency requirements. Since 2 August 2026, key provisions of the EU Artificial Intelligence Act have begun to apply, including new rules designed to help people recognise when they are interacting with AI or encountering content generated or altered by AI.
The objective is straightforward: people should know when technology is shaping what they see, hear or read.
Under Article 50 of the AI Act, certain AI systems must make it clear to users when they are interacting directly with an AI system. Providers of systems capable of generating or manipulating content must also enable appropriate technical identification of AI-generated material.
Particular attention is given to deepfakes—AI-generated or manipulated images, video and audio that can make people appear to say or do things that never happened.
The rules also introduce transparency requirements for certain AI-generated or manipulated text concerning matters of public interest, particularly when such content has not undergone human review or editorial control.
For citizens, the principle is important: AI should not operate invisibly when its use can affect our understanding of reality.
The rapid development of generative AI has made misinformation easier to produce and potentially much harder to detect. A fabricated image of a public figure, a synthetic video or an AI-generated article can circulate widely before anyone has had the opportunity to verify it.
Transparency can therefore become an important layer of protection.
But labelling AI content is not a complete solution. People also need the ability to interpret those labels, question information and verify sources. This makes digital and AI literacy increasingly important—not only for young people, but for society as a whole.
Not everyone sees stronger regulation as an uncomplicated solution.
Technology companies and startups operate in a highly competitive global environment, and additional regulatory requirements can increase compliance costs and complexity. There are concerns that excessive regulation could make Europe less attractive for some AI companies or slow experimentation and innovation.
This creates a difficult policy balance.
Regulation that is too weak may leave citizens exposed to manipulation, fraud and misinformation. Regulation that is unnecessarily restrictive could discourage innovation and make it harder for European organisations to compete in a rapidly developing technological sector.
The challenge is therefore not simply to regulate AI—but to regulate it intelligently.
One of the most important lessons from the debate is that regulation cannot replace human judgment.
Even when content is properly labelled, citizens still need to ask: Who created it? Why was it created? Is the information credible? What evidence supports it?
For educators, civil society organisations and institutions working with young people, this represents a major opportunity. AI literacy should not mean simply learning how to use AI tools. It should also mean understanding their limitations, recognising manipulation and making responsible decisions about information.
The future of AI will therefore depend on two things developing together: technological innovation and human capacity to use technology critically.
The question is not whether AI should be regulated. The more important question is how we can create rules that protect people without preventing responsible innovation.
That conversation has only just begun.
Source: Adapted and translated from the article “Nova pravila za AI – Zaštita od manipulacija ili kočnica za inovacije?”, published by Mreža mira on 19 August 2026.