RGAA 5 - A Welcome Update, But Questions Remain

The French Interministerial Directorate for Digital Affairs (Dinum) announced in early March 2026 that it was working, alongside public, non-profit, and private stakeholders, on drafting RGAA 5, the next version of the General Accessibility Improvement Framework (Référentiel Général d'Amélioration de l'Accessibilité).

Expected publication: late 2026. The topic deserves a closer look, with a slightly less institutional perspective than the official press releases.

What is the RGAA, and why does it matter?

The RGAA is the framework tied to French law that sets digital accessibility obligations for public services and large companies.

The current version, 4.1.2, is based on the W3C's WCAG 2.1. It defines concrete criteria that developers and designers must meet to ensure their interfaces are usable by people with disabilities.

This is not an abstract issue: 12 to 15 million people in France live with a disability, and the law has required digital accessibility since 2005. Twenty years later, compliance levels remain woefully insufficient (roughly 40 fully compliant sites out of a sample of 7,000+ websites), even across the majority of public services. That is the context in which RGAA 5 arrives.

What is actually changing

Alignment with WCAG 2.2

The most significant change is the move from WCAG 2.1 to WCAG 2.2, published in December 2024 by the W3C. In practice, this introduces several new or strengthened criteria.

• Focus visibility: three criteria clarify the requirements. The focus indicator must not be hidden by other page elements (2.4.11, level AA), with an even stricter requirement at AAA (2.4.12), and minimum size and contrast constraints for the focus indicator itself (2.4.13, AA). A useful improvement, as focus handling is often the most overlooked aspect of accessibility reviews.

• Interactions: drag-and-drop gestures must now offer an alternative accessible via a single pointer (WCAG 2.5.7, AA). A point frequently forgotten in modern interfaces, particularly in content management tools or Kanban boards.

• CAPTCHAs: cognitive tests (puzzles, memory challenges, etc.) may no longer be required in authentication processes (WCAG 3.3.8 and 3.3.9) or other types of forms. Great news, as CAPTCHAs are notoriously inaccessible for many user profiles. (And based on those CAPTCHAs, I'd apparently be classified as a robot myself. 😁)

• Target sizes (click, touch): a minimum size of 24x24 CSS pixels is set (WCAG 2.5.8, AA). With wording that leaves room for interpretation, modest, but better than nothing.

• Interface consistency: help mechanisms such as contact options or FAQs must appear consistently across all pages (WCAG 3.2.6, A). Again, an angle that is frequently neglected.

• Forms: information already entered by the user within a process must not be asked again without good reason (WCAG 3.3.7, A). Common sense and accessibility aligned for once.

Coverage of new content types

RGAA 5 should, for the first time, define technical criteria for mobile applications and office documents (PDF, Word, etc.), something the Luxembourg Information Press Service already does with RAWeb, RAAM, and RAPDF. This is a genuine step forward. These formats make up a significant share of exchanges between public services and users, yet they have been largely absent from the formal scope until now. Private companies will no longer be able to hide behind inaccessible PDFs and mobile apps either.

Institutional changes

On the legal side, two changes are planned: the designation of Arcom (the audiovisual and digital communications regulatory authority) as the supervisory body, and the introduction of a dedicated online service for submitting and publishing accessibility statements. The latter is welcome: today, tracking accessibility declarations remains poorly tooled and hard to follow.

What still needs clarifying

The official announcement is deliberately cautious, which is understandable for a version still being drafted. But a few grey areas are worth flagging.

• The rewording of criteria is presented as a simplification goal. That is positive in theory, but rewording can also introduce new ambiguities. Without seeing the text, it is impossible to judge. Worth watching closely during the public consultation. From what I understand, the question-based format may be dropped, though that information still needs confirming.

• The timeline remains vague. "Late 2026" is a target, not a firm date. Yet public digital projects need visibility. A late or delayed publication risks creating an uncomfortable period of uncertainty for teams currently working toward compliance.

• Enforcement remains the real question. The RGAA has existed for years, and so have the legal obligations. Yet compliance rates remain very low. The arrival of Arcom as a supervisory authority is presented as progress, but its actual capacity to process complaints and issue sanctions remains unproven. Digital accessibility suffers less from a lack of rules than from a lack of effective enforcement.

• The dependency on AG remains, in my view, worth questioning. Given how slowly that framework evolves relative to the pace of change in actual usage, and the fact that it took the RGAA two years to catch up, this creates what I see as a structural drag on standards development.

Should I wait for RGAA 5, or can I still use RGAA 4.1.2?

No need to wait, and Dinum says so with commendable clarity: ongoing work based on RGAA 4.1.2 must not be suspended. Several strong arguments support this.

First, RGAA 5 criteria will refine and complement those of RGAA 4.1.2, not invalidate them. The differences between WCAG 2.1 and 2.2 are real but limited. What is done correctly today will remain correct tomorrow.

Second, accessibility statements published before RGAA 5 is released will remain valid for 18 months after its publication (within a three-year limit from their original date of issue). No brutal reset.

Finally, and perhaps most directly: every month without compliance excludes users. Waiting for a new version of the framework is not an acceptable argument in the face of that reality.

It is even possible to get ahead by incorporating WCAG 2.2 criteria not yet covered by RGAA 4.1.2 into current projects. That is the most coherent approach for serious teams. (Something you can already do using the Custom Guidelines feature in CheckFox.)

CheckFox take

RGAA 5 is a logical and long-awaited update that addresses real gaps (mobile apps, documents, WCAG 2.2). Aligning with international standards is a good thing, and the push to simplify criteria could make them easier to adopt for non-specialist teams.

But digital accessibility in France needs less a better framework and more a better culture around applying it. The tools exist. The legal framework exists. What is often missing is dedicated human resources, allocated budgets, and a genuine willingness to treat accessibility as a non-negotiable constraint rather than a box to tick.

The real question posed by RGAA 5 is not "what new criteria?" but "how do we finally move from surface-level compliance to actual accessibility?". That is where the coming months will be telling.

Sources:

• Dinum announcement - 2 March 2026
• Adimeo - A new version of the RGAA for 2026*