Charter Review

Heather: I had a query which I posted in Slack. We were hesitant to take this on becuase we know it will go through the process and raise formal objections so I didn't put any official comments anywhere but I've was wondering about AI and Agentic browsers potetntially expands UA definition, is there a similar thread to pull when talking about AI enabled accessibility tools? Who is responsible for accessibility? e.g. if they're not only reading the screen but also filling out information and doing stuff with the content?

Yves: Also for the conformace rule, it might be linked. Is a site conformant if instead of following WCAG, it relies on AI to make it conformant.

Heather: I'm hearing about website written for AI not for humans. This would be the accessiblity side of that.

Lola: Your question about AI accessibility tools and who is responsible is definitely worth bringing up to the group if it's not in the charter. It may be something we think they should think about. They might say it's not in their remit. I vaguely remember that WCAG 3.0 is going to be thinking about the role of AI. Maybe they are already thinking about it, but they aren't ready to formalize it in the charter. But it is valid to raise regardless.

Jeffrey: I've been interacting with this a lot from the Google side, but I think TAG doesn't need to make a comment; they are not architectural issues.

Lola: Some of the issues I've heard are around the work they're committing to and the timeframe they're committing to, to do it. It's not architectural, but if they don't deliver, that could have impact. There are a number of people who don't think they're going to deliver to this timescale.

Jeffrey: What could we say about the charter that would help them move faster?

Lola: I think we could encourage them to do less, but I'd have to think about it.

Matthew: I noticed one significant thing that they are commiting to doing less of: 2.x maintenance. This has been a tough balancing act for them: there's a lot of work for 2.x, but 3 is going well. They've said, "We're not going to do anything on 2.x in this charter period, but if we learn that there is more work needed there, then we will invest time for that in the future." I think it's great that they've scoped it in that way.

Jeffrey: Another aspect that reassured me: some people have said it won't be ready until 2030, but the charter says CR in 2028. But a few experienced people talked about how the group tends to spend time in CR to get wide feedback, so that reassures me that the two timescales are compatible.

Lola: So seems not much for TAG to say here. Other avenues for other comments.

Jeffrey: We should decline this perhaps - will draft a comment and run it by Heather.