Skip to main content
v2026.11,986 entries · CC-BY 4.0

Editorial · CASRAI · AI and ML research outputs

EU proposes Digital Fairness Act with AI consumer safeguards

The EU proposed AI purchasing controls and digital fairness safeguards on November 11, with specific education and science exclusions.

Published 14 Nov 2026· 3 minute read

Ask CASRAI · free to try

Ask about this story

Ask your first 2 questions free below. Subscribers get 150 a day for $29 a month.

An AI assistant specialized in research administration. It cites the sources behind every answer, labels web answers and says when it can't answer.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays free to read, with no account and no card.

CASRAI is the reference for research administration — bookmark it for the next question.

The European Commission proposed a Digital Fairness Act on November 11, 2026, including safeguards for AI purchasing agents and digital features that encourage compulsive use. The proposal now goes to the European Parliament and Council for discussion; these proposed safeguards are not enacted requirements. The Commission’s announcement places it within a broader consumer protection package.

The Commission’s proposal page also identifies unfair personalisation, subscription cancellation and access to human customer service among the problems the package addresses.

What the proposed AI safeguards cover

The draft regulation’s relevant provisions are compared below. The research questions are editorial suggestions for evaluation, rather than proposed legal duties.

Proposed provision What it would do Research question
Article 17: addictive design Require effective controls to disable features encouraging compulsive or excessive use, including persistent automated prompts simulating relationships, subject to exclusions. Can users stop relationship-based engagement pressure?
Article 26: purchasing agents Require adjustable approval settings. Consumers could confirm all, selected or no contracts; paid recommendations or conflicts would always trigger disclosure and confirmation. Does the agent honour approval settings and expose commercial conflicts?
Articles 12–13: personalisation Restrict profiling-based advertising using known minors’ or sensitive data, with a DSA platform exclusion; restrict personalised pricing likely to harm known minors or vulnerable consumers. Does testing distinguish targeting, pricing, vulnerability and harm?

Where university research and spinouts fit

Articles 2–3 frame the proposal around business-to-consumer commercial practices. Article 17 specifically excludes the education and science categories referenced in Article 2(4)(c) and (e) of the proposed EU KIDS Act.

Those EU KIDS Act provisions cover services and systems designed primarily for education and operated by educational establishments or organisations, or on their behalf; and services and systems developed and operated solely for scientific research and development. These are the categories cross-referenced by the DFA’s addictive-design exclusion, rather than an exemption from the entire DFA for every university activity.

For researchers, the comparison suggests concrete additions to a model evaluation suite: examine whether users can disengage, whether purchasing limits hold, and whether commercial incentives are visible. These are potential study questions, not evidence that any particular assistant has failed them.

For research offices and technology transfer teams, the implication is to examine a project’s purpose, operator and commercial use before assuming an exclusion applies. A consumer-facing spinout merits its own scope assessment; university origins alone do not establish that the specific Article 17 categories fit.

How the proposal relates to the AI Act and DSA

The Commission’s Consumer Package Q&A describes the DFA as a consumer protection safety net complementing the existing digital rulebook. The EU AI Act provides a separate framework based on AI risk, while the DSA addresses platform responsibilities. The DFA proposal would reach businesses engaged in consumer commercial practices beyond platforms, including through its personalisation provisions.

The Q&A also distinguishes this broader consumer framework from the proposed EU KIDS Act’s focus on minors, including users of AI companions and conversational chatbots. Researchers assessing persuasive assistants therefore have several distinct policy instruments to track, each with its own scope.

Legislative next steps and the separate enforcement proposal

The Commission presented a separate revision of the Consumer Protection Cooperation Regulation alongside the DFA. That enforcement proposal would give the Commission powers to investigate and fine businesses in large cases affecting consumers across the Single Market. Those proposed powers should be distinguished from the DFA’s substantive consumer safeguards.

The Commission says the proposals will now be discussed by Parliament and Council. Its announcement supplies no adoption deadline. For university product teams, the immediate significance is a published legislative proposal to assess and follow as negotiations proceed, rather than a new implementation date to enter in a compliance calendar.

Sources

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →