AI Act Guide

If you have an AI chatbot on your website, visitors must know they're talking to a computer. This is one of the most common AI Act obligations for webshops — and luckily one of the easiest to fix.

Using AI for product descriptions, blog posts or images? Then you must disclose it. The AI Act requires both visible labels and technical metadata.

The AI Act bans certain AI applications outright: social scoring, subliminal manipulation, emotion recognition in the workplace and scraping faces for databases. The fine: up to 7% of annual turnover.

Article 4 obliges organisations to give staff who work with AI sufficient knowledge. It has been in force since February 2025 and is the most underestimated obligation in the AI Act.

Does your webshop show 'recommended products' or 'others also bought'? The AI Act and the Digital Services Act both require you to explain how these systems work.

AI for credit scoring, insurance, HR decisions or recruitment is high-risk. The obligations are heavy: conformity assessment, logging, human oversight and impact assessments.

AI-driven pricing is not banned, but it becomes so when it exploits vulnerable groups or manipulates people unconsciously. Transparency about how prices are formed is always required.

If you use AI for tracking, personalisation or search results, you must specifically disclose this in your cookie banner and privacy policy. 'Marketing cookies' alone is not enough.

Emotion recognition in the workplace is prohibited. On customers it is high-risk: only allowed with transparency and an opt-out. The difference is the context.