Short answer: GDPR protects personal data, whatever technology processes it. The EU AI Act regulates AI systems, whether or not they touch personal data. A website chatbot is a good example of both at once: the AI Act says visitors must be told it’s an AI, and GDPR says you need a legal basis and a privacy notice for the chat transcripts it stores. You have to comply with both, and meeting one doesn’t satisfy the other.
The two laws side by side
| GDPR | EU AI Act | |
|---|---|---|
| Protects | People’s personal data | Health, safety and fundamental rights affected by AI |
| Applies to | Any processing of personal data | AI systems and general-purpose AI models |
| Main roles | Controller, processor | Provider, deployer (plus importer, distributor) |
| Approach | Principles plus rights for individuals | Risk tiers: banned, high-risk, transparency, minimal |
| Applies since | 25 May 2018 | In stages, 2025 to 2028; transparency rules since 2 August 2026 |
| Max fines | 20 million euros or 4% of turnover | 35 million euros or 7% for banned practices; 15 million or 3% for most other breaches |
| Enforced by | Data protection authorities | Market surveillance authorities (in Germany mainly the Bundesnetzagentur) |
| Individual rights | Access, erasure, objection, portability and more | Complaint to an authority; explanation of some high-risk decisions |
Roles don’t line up neatly
A website owner using a chatbot widget is typically the GDPR controller of the chat data and the AI Act deployer of the chatbot. The widget company is usually your GDPR processor and the AI Act provider. But the mapping isn’t automatic. A vendor that uses your visitors’ chats to train its own models acts as a controller for that purpose, which needs its own legal basis and transparency. Check the vendor’s terms.
Where the two laws overlap on a website
Chatbots
- AI Act: tell visitors they’re talking to an AI at the first interaction (Article 50(1)).
- GDPR: privacy notice naming the vendor, purpose and retention; data processing agreement; legal basis for storing transcripts; consent if the widget sets non-essential cookies.
More in our chatbot compliance guide.
Automated decisions
GDPR Article 22 gives people the right not to be subject to decisions based solely on automated processing that have legal or similarly significant effects, with exceptions, and requires meaningful information about the logic involved. The AI Act adds obligations if the system is high-risk, for example AI that screens job applicants or assesses creditworthiness. A website form that uses AI to reject loan or job applications automatically triggers both.
Personalisation and profiling
GDPR covers the profiling itself: legal basis, transparency, the right to object. The AI Act generally doesn’t regulate recommendations on a normal website. Consumer law adds a duty to disclose personalised prices. See our ecommerce guide.
AI-generated content
Labelling AI text and deepfakes is an AI Act matter (Article 50(4)). GDPR comes in when the content is about real, identifiable people: a deepfake of a real person is also processing of their personal data and can breach GDPR on its own.
Biometrics
Emotion recognition and biometric categorisation trigger an AI Act notice duty (Article 50(3)) and, in some settings, an outright ban. Biometric data used to identify people is special-category data under GDPR, with strict conditions.
What only GDPR asks for
- A legal basis for every processing of personal data.
- Privacy notices, data subject rights, data processing agreements.
- Data protection impact assessments for high-risk processing.
- Rules on transfers outside the EU.
What only the AI Act asks for
- Telling people they’re interacting with an AI, even when no personal data is processed.
- Labelling deepfakes and unreviewed public-interest AI text.
- Machine-readable marking of generated content (providers).
- AI literacy for staff using AI (Article 4).
- Technical requirements, conformity assessment and registration for high-risk AI.
A combined checklist for website owners
- List every AI tool on your site and what data it touches. Our website AI audit has a template.
- For each tool: AI Act duty (if any), GDPR legal basis, vendor role, DPA in place.
- Update the privacy notice with AI vendors, purposes and retention.
- Add AI disclosures where Article 50 requires them. See AI transparency rules for websites.
- Turn off vendor training on your data unless you have a basis for it.
- Keep both records together, because authorities from both sides may ask.
The free EU AI Act Ready plugin covers the AI Act disclosure side on WordPress and runs on your own server without sending visitor data anywhere, so it doesn’t add a new processor to your privacy notice.
Frequently asked questions
If I’m GDPR-compliant, am I AI Act-compliant?
No. GDPR doesn’t require AI disclosure, deepfake labels or AI literacy. You need both.
Can I be fined under both laws for the same thing?
The two laws protect different interests and are enforced by different authorities, so one incident can breach both. The AI Act asks authorities to consider fines already imposed for the same conduct when setting the amount.
Does the AI Act apply if my AI tool processes no personal data?
Yes. A chatbot that stores nothing still has to say it’s an AI.
Which law is stricter?
For a normal website, GDPR has more day-to-day obligations. The AI Act’s duties for websites are narrower but they’re new, and most sites haven’t implemented them yet. See what the EU AI Act means for your website.