From 2 August 2026 the transparency rules of the EU Artificial Intelligence Act apply in full. For most businesses that means one simple thing: if customers talk to an AI — in chat, on Viber or by phone — they must know it's an AI.
This guide is a practical summary, not legal advice. For a specific case, consult a lawyer.
What the AI Act is
Regulation (EU) 2024/1689 is the EU's first general law on artificial intelligence. It entered into force on 1 August 2024 and applies in stages. As a regulation it applies directly in Bulgaria, with no separate national law needed.
The key dates
| Date | What applies |
|---|---|
| 2 February 2025 | Prohibited AI practices and the AI literacy duty for staff (Art. 4) |
| 2 August 2025 | Obligations for providers of general-purpose AI models |
| 2 August 2026 | General application, including transparency rules (Art. 50) and penalties |
| 2 December 2027 | High-risk systems under Annex III — postponed by the AI Digital Omnibus |
The delay covered high-risk systems but not the chatbot transparency rules. Systems placed on the market before 2 August 2026 got a short transition only for machine-readable marking of generated content.
What Article 50 requires
- Chatbots and voice assistants — people must be informed they're interacting with an AI, unless it's obvious from context (Art. 50(1)). This duty sits with the provider who designs the system.
- Generated content — systems producing synthetic text, images, audio or video must mark it in a machine-readable format (Art. 50(2)).
- Emotion recognition and biometric categorisation — people must be told (Art. 50(3)).
- Deepfakes — realistic generated images, audio or video must be disclosed. AI-generated text published on matters of public interest too, unless a human with editorial responsibility reviewed it (Art. 50(4)).
Who is responsible — you or your supplier
The AI Act separates the provider (who develops a system or places it on the market under their name) from the deployer (who uses it in their business). The duty for a chatbot to say it's an AI sits with the provider. But if you commission a chatbot and launch it under your brand, you may end up as the provider. The safest route is to write the requirement into the contract — and have the system built that way from the start.
The fines
Breaching the transparency rules can cost up to €15m or 3% of worldwide annual turnover, whichever is higher (Art. 99(4)). For SMEs, the lower of the two applies (Art. 99(6)).
A 7-point check
- The chatbot opens with something like “Hello, I'm the AI assistant of …”.
- The voice assistant says it at the start of every call.
- There's an easy way to reach a person.
- Realistic AI images and videos in your ads are labelled.
- Your privacy policy mentions AI processing.
- Your supplier contract covers Art. 50 and GDPR compliance.
- Staff who use AI have been trained (Art. 4) — and it's documented.
How we handle it
Every chatbot and voice assistant we build introduces itself as an AI, has a button or command to reach a person and keeps conversation logs for the agreed period. We sign a data processing agreement and document compliance in the handover pack.
Frequently asked questions
Does my chatbot have to say it's an AI?
Has the AI Act been postponed?
What are the fines for a small business?
Sources
- Regulation (EU) 2024/1689 — EUR-Lex
- European Commission — Guidelines on transparency obligations under Art. 50
- Ropes & Gray — Operationalising the EU AI Act's transparency obligations (2026)
- Facephi — The AI Act enters a new phase (2026)
Want the numbers for your business? The free consultation takes 30 minutes.
Book a time