The EU AI Act Enters Its Main Phase: What Changes From 2 August
The EU's AI rulebook reached its main application date on 2 August 2026. Chatbots must now disclose themselves, deepfakes must be labelled - and the rules reach any company selling into Europe, from Istanbul to Tashkent.
On 2 August 2026, the European Union's Artificial Intelligence Act reached its main application date. After a staggered rollout that began with outright bans on practices such as social scoring in early 2025 and transparency duties for general-purpose models later that year, most of the law's day-to-day rules are now enforceable across all 27 member states. For users, the change will be visible in small but telling ways. For companies far beyond Europe's borders — including in Türkiye and Central Asia — it will be visible on their compliance checklists.
What actually changed on August 2
The most immediate obligations concern transparency. Chatbots and other interactive AI systems must now tell users, clearly, that they are dealing with a machine rather than a human. Providers of systems that generate synthetic audio, images, video or text must ensure their outputs are detectable in a machine-readable form. And anyone deploying deepfake technology must disclose when content has been artificially generated or manipulated, according to the European Commission's enforcement notice.
These labelling duties arrive at a moment when synthetic media has become a central concern for governments and platforms alike. The EU's answer is not to ban the technology but to make its origin traceable — a regulatory philosophy that will now be tested at scale.
Not everything applies at once
The Commission's timetable remains deliberately staggered. Stand-alone high-risk systems listed in Annex III of the Act — categories such as recruitment, credit scoring and law enforcement — have until 2 December 2027 to comply, while AI embedded in products already regulated under EU law, from medical devices to vehicles, has until 2 August 2028. Compliance, however, must increasingly be demonstrated in practice rather than on paper, and regulators now face scrutiny of their own: whether 27 national authorities can enforce a single rulebook consistently.
The stakes are considerable. Breaches of the Act's prohibited practices can draw fines of up to 35 million euros or 7 percent of global turnover, whichever is higher.
Why this matters beyond Europe
The AI Act applies to any provider placing an AI system on the EU market, regardless of where that provider is established. That extraterritorial reach — the familiar "Brussels effect" that turned GDPR into a global standard — is the part of the story that matters most for Türkiye and Central Asia.
Turkish software houses selling AI-enabled products into Europe, Kazakh and Uzbek platforms courting European users, and regional startups pitching EU investors will all find the same questions waiting: Is your chatbot disclosed? Is your synthetic content marked? Can you document how your model was trained? For exporters, EU compliance is becoming less a legal burden than a market-access badge — proof of trustworthiness that travels well beyond the single market.
The region's governments are watching too. As we reported from the Shusha Global Media Forum, where President Aliyev set out his vision of AI and trust, and as capitals from Astana to Tashkent accelerate their digital agendas, the EU's rulebook is the reference point most regional strategies quietly measure themselves against. The question of technological sovereignty — who sets the rules, and who merely follows them — runs through every one of these debates.
The transparency bet
Europe is wagering that trust, not speed, will decide the AI race. Critics argue the Act risks slowing European innovation while the United States and China sprint ahead; supporters counter that calibrated transparency is precisely what gives leaders room to act. What is no longer in dispute is that the experiment has begun in earnest. From this month, an AI system operating in Europe must say what it is — and prove what it does.
Oğuzhan Kuri
Software Director at EUREFLECT, responsible for the technology behind the organization's digital platforms. He also writes on artificial intelligence, digital regulation and the technology politics of Europe and Eurasia — where engineering meets policy.
