Use-case guide

AI in Retail & E-commerce

Retail AI is becoming a regulated practice in three areas: dynamic pricing (algorithmic price discrimination), recommendation algorithms (DSA scrutiny + dark-pattern enforcement), and AI-driven customer service (Article 50 disclosure + state chatbot laws). The shift since 2023: regulators stopped treating algorithmic pricing as a competition-only question and started treating it as a consumer-protection question. The FTC's 'surveillance pricing' inquiry (2024) and the EU's DSA-driven dark-pattern enforcement set the new baseline. Add California SB 243 chatbot disclosure, Connecticut AB 6691, and similar — retail is suddenly multi-jurisdictionally regulated for the first time.

For: E-commerce platform owners, marketplace operators, pricing/recommendation engineers, consumer-protection counsel, DTC brand technology leads

What's at stake

Dynamic pricing AI is the new consumer-protection frontline

FTC's 6(b) Surveillance Pricing Inquiry (Jul 2024) and follow-up reports are mapping the practices. State AGs (CA, NY, MA) are subpoenaing pricing-model documentation. Individual-targeted pricing using AI inferences is increasingly treated as 'unfair' under state UDAP statutes.

DSA recommendation-system rules apply to retail platforms

Marketplaces qualifying as VLOPs (Amazon, eBay, AliExpress, Booking — all designated) must offer at least one non-profiled recommendation option; transparency on the main parameters; user opt-out from profiling-based personalisation.

Chatbot disclosure laws across multiple states

California SB 243 (2024-2025), Connecticut AB 6691, NJ pending — require disclosure when consumer is interacting with an AI chatbot. EU AI Act Article 50 globalises the duty.

Dark-pattern enforcement intensifying

FTC's 2022 dark-pattern report + 2023 enforcement actions (Epic, Vonage, Click-to-Cancel rule) — AI-driven nudges, urgency, and personalised pressure tactics are now in the FTC's active enforcement lane.

Regulations that apply

EU AI Act

Law
→

Article 50 chatbot disclosure; Article 5(1)(a) prohibition on manipulative AI; Article 5(1)(b) exploitation-of-vulnerability ban (relevant for vulnerable-consumer targeting).

Where in the text: Articles 5(1)(a), 5(1)(b), 50.

EU Digital Services Act

Law
→

Recommendation-system transparency (Article 27); VLOP risk assessment (Article 34); ban on profiling-based ads for minors (Article 28); dark-pattern prohibition (Article 25).

Where in the text: DSA Articles 25, 27, 28, 34.

California SB 243 + chatbot disclosure laws

Law
→

Disclosure required when consumer interacts with an AI chatbot influencing purchase decisions. State patchwork emerging: CT, NJ, MA following.

Where in the text: Cal. Bus. & Prof. Code § 17941 (proposed/passing).

FTC Act Section 5 + Section 6(b) Surveillance Pricing Inquiry

Law
→

FTC's authority to challenge algorithmic pricing as unfair / deceptive. 2024 6(b) inquiry results expected to drive rulemaking or follow-on enforcement.

Where in the text: 15 U.S.C. § 45; FTC 6(b) Surveillance Pricing Order (2024).

Do

  • ✓Disclose AI use in personalisation prominently — the 'pre-purchase information' duty under DSA + state laws expects it before, not after.
  • ✓Maintain non-profiled-recommendation toggles per DSA Article 27 for any feature that ranks products algorithmically.
  • ✓Document pricing-model inputs explicitly — FTC + state-AG subpoena patterns target the exact feature set used.
  • ✓Apply chatbot AI-identification at the START of conversations, not when asked. EU AI Act Article 50 wording is on this point.
  • ✓Audit recommendation models for protected-class disparate-impact (price + product visibility). Several state AGs now use this lens.

Don't

  • ✗Don't price-discriminate based on inferred protected-class membership. This is the lane that triggers civil-rights + UDAP joint enforcement.
  • ✗Don't deploy 'urgency' or 'scarcity' messaging that's algorithmically-generated and false — FTC dark-pattern enforcement covers this directly.
  • ✗Don't use AI emotion-recognition in retail interactions in the EU. Article 5(1)(f) restrictions extend beyond workplace + school to specific consumer-vulnerability cases.
  • ✗Don't withhold the non-profiled-recommendation toggle from EU users — DSA enforcement against this is active.
  • ✗Don't bypass the chatbot disclosure with copywriting tricks ('Hi, I'm Sarah from customer service' when Sarah is an AI). California SB 243 wording captures the practice.

Also worth knowing

For grocery and food retail: dynamic pricing intersects with state weights-and-measures laws (price-tag-vs-shelf-tag mismatches) — that's the consumer-protection lever AGs use even when AI pricing isn't named. For travel and ticketing platforms: pricing-AI scrutiny is intensifying; UK CMA + Australian ACCC pursuing 'drip pricing' cases. For B2B marketplaces: most consumer-protection rules don't apply, but DSA still does because the platform-vs-user relationship triggers it independently.

Want a tailored answer?

The wizard takes your jurisdiction, AI use case, and data types and gives you the top-3 regulations to focus on — in 60 seconds.

Start the wizard →

Educational guide. Not legal advice. For specific compliance decisions, consult qualified counsel in the relevant jurisdiction.

Note: this guide was drafted with AI assistance — Anthropic Claude.