> "Les modèles d'#IA constituent une nouvelle couche, différente" sur #internet. "Ils utilisent le fait que le web contient énormément de données pour être entraînés", pointe M. Berners-Lee pour qui il s'agit d'une innovation "passionnante" mais qui gagnerait à être encadrée.
Il souligne notamment que si le web dispose du #W3C (World Wide Web Consortium), organisation de normalisation d'internet, l'IA ne dispose pas d'un organisme collaboratif équivalent.
En effet
The W3C has officially rechartered the Dataset Exchange Working Group and @fracoparis will be chairing this new iteration.
The main objective is to develop a W3C Recommendation for documenting datasets at the variable level. The group will also revise existing vocabularies like Data Cube and DCAT. See the new charter for more details.
<p>The mission of the Dataset Exchange Working Group is to:</p> <ul> <li>Develop a W3C Recommendation for the interoperable exchange on the Web of information on datasets at the variable level (Variable Description Vocabulary, VVD).</li> <li>Maintain and revise the RDF Data Cube Vocabulary, QB, taking into account feature requests from the QB user community.</li> <li>Maintain and revise the Data Catalog Vocabulary, DCAT, taking into account feature requests from the DCAT user community.</li> <li>Define and publish guidance on the specification and use of application profiles when requesting and serving data on the Web.</li> </ul>
@abucci Me (a person with Myopia) stumbled over such problem with selecting and copying by Ctrl+C on websites where content was hidden and only visible for a screen reader.
Pasting in a app resulted in more elements i had selected.
I thought that was a bug in my browser or Windows clipboard and tried multiple times, thinking myself is dumb to use a website.
But as a Webdeveloper i was able to inspect the website and found the culprit.
I contacted the webmaster of the page and told them to think about better accessibility and begged for a fix.
Myself as a webdev pro learned hiding content for screen readers by stylesheet markup can cause more trouble.
Yes, and i agree in your mention that such is causing cognitive trouble and is poor webdesign .
Unfortunatly myself has not enough knowledge to interpret #W3C #WAI and #WCAG standards to others a better solution.
And i guess on my private sites i made some errors which may cause cognitive trouble for some other people – makes me feeling ashamed.
But we are learners, even when we were/are disability activists for a long time.
#W3C initiates leadership transition
https://www.w3.org/news/2026/w3c-initiates-leadership-transition/
> W3C leadership transitionW3C has announced a leadership transition. Dominique Hazaël-Massieux has been appointed Interim CEO with immediate effect.
💥
Grounding AI the webby way — Taxonomy Boot Camp London 2026 takeaways
There was a heavy focus on artificial intelligence. But what really struck me was that the semantic approaches that are preparing us for our AI future are well-established web standards that have been around for decades.https://duncanstephen.net/grounding-ai-the-webby-way-taxonomy-boot-camp-london-2026-takeaways/
Really angry about #W3C. They provide great tools like #HTML <datalist> ... and make it non-styleable! The same error they made with <select> repeated 20 years later. In the end, developers have to dump a ton of JS on that and ruin a11y just because the design requires a different background color.
Have the people who came up with this ever bothered to build a website or an app?
out of bounds, anywhere out
PPC Land
Making Sense
DoctorG ♀️🏳️🌈
W3C Developers
Wladimir Mufty
Karl Dubost
Duncan Stephen
Coralie Mercier (W3C)
wortwart