The EU’s Digital Fairness Act Wants To Take On Online Manipulation And AI – What Would Change For Consumers?

Trying to cancel a digital subscription often feels like trying to break out of Alcatraz. Joining takes seconds, but unsubscribing involves endless confirmation pages and emotional blackmail about abandoning progress. The EU appears determined to dismantle the trap.

A leaked draft of the Digital Fairness Act, set to be unveiled by the European Commission on Wednesday 11 November, would replace the ageing 2005 Unfair Commercial Practices Directive with a binding regulation. The legal difference carries weight, because directives get diluted through national legislation, whereas regulations apply identical rules across all 27 member states at the same time.

Those 2005 rules arrived long before dark patterns, subscription traps and AI-fuelled retail tactics became standard practice. This draft is the EU trying to catch up with the internet it’s meant to police, though tough negotiations between the European Parliament and member states lie ahead before anything reaches the statute book.

Ten Ways To Annoy A Shopper, Now Up For Banning

The draft reportedly blacklists ten dark patterns, meaning interface designs that distort how people make economic decisions. Each would be treated as unfair in all circumstances, so regulators wouldn’t need to show consumer harm case by case.

The first category zeroes in on exit routes and sneaky pricing. Making cancellation disproportionately harder than signing up would be banned, alongside burying prices or auto-renewal terms and falsely suggesting that no alternative exists.

Next up is artificial panic. False or unqualified claims about availability or popularity would be tossed out, which should retire those infamous “only two left at this price” alerts, and repeated scarcity warnings during checkout would go with them. Pre-selected add-ons and extra quantities face the chop as well, so shoppers only pay for what they pick themselves.

The remaining four rules focus on badgering and deliberate confusion. The draft would ban framing a refusal as the wrong choice, repeatedly pressing someone to reverse a decision, confusing double negatives and fake notifications that imply someone has been in touch.

To seal the exits, the draft reportedly requires a “terminate the contract here” button for online subscriptions, catching free trials before they turn into permanent bank drains. Sellers would also have to show the maximum price a customer may pay before checkout, which targets drip pricing, where mandatory fees appear at the finish line.

The Chatbot Gets An Off Switch And The Shopping Agent Gets A Leash

The AI provisions are where the draft breaks new ground. Users would get the right to switch off chatbot functions that simulate interpersonal relationships in order to prolong engagement. Companion-style chatbots wouldn’t be banned outright, but adult users could turn those features off.

AI shopping agents, which browse, compare or buy on a consumer’s behalf, would get their own rules. The draft reportedly requires them to disclose sponsored suggestions and lets users decide which purchases need explicit approval. It would be the first time EU consumer law sets dedicated rules for the category.

The response cleanly addresses an obvious point of friction. An AI agent that nudges you towards whoever paid for the placement is just an advert in a better suit.

Fines Big Enough To Reach The Boardroom

The real bite of the draft lies in its enforcement powers. It reportedly creates an EU-wide “intra-Union” infringement category for widespread breaches, with fines of up to 6% of a trader’s worldwide annual turnover. Numbers like that turn dark patterns and subscription traps into a board-level liability.

The leaked text could still change once the Commission publishes its formal proposal. If adopted, the draft is reported to apply 12 months after entry into force, and legal analysis doesn’t expect the Act to take effect before the end of 2028 at the earliest, given that Parliament and the Council both need to agree.

The details to watch in negotiation are how precisely the blacklist survives and whether regulators get the resources to enforce a 6% penalty across borders. The rules for AI agents will be tested too, as those tools become the go-between for more purchases. The Commission’s proposal is a starting point, and the version that reaches the statute book may look quite different.