HTTP/2 200 date: Wed, 08 Oct 2025 02:06:48 GMT content-type: text/plain; charset=utf-8 content-length: 4125 cf-ray: 98b214bbea63c1c4-BLR content-location: 13-rqtf-irc.txt vary: negotiate,accept,accept-charset,Accept-Encoding,Origin tcn: choice last-modified: Wed, 13 Nov 2024 15:19:19 GMT etag: "287c-626ccdb8e8bc0;62e1ac79982b4 cache-control: max-age=21600 expires: Wed, 08 Oct 2025 08:06:48 GMT content-encoding: gzip access-control-allow-origin: * x-backend: www-mirrors x-request-id: 98b214bbea63c1c4 strict-transport-security: max-age=15552000; includeSubdomains; preload content-security-policy: frame-ancestors 'self' https://cms.w3.org/ https://cms-dev.w3.org/; upgrade-insecure-requests cf-cache-status: BYPASS accept-ranges: bytes set-cookie: __cf_bm=DHNz5wqX11ZbAaFC.0CClHI3MAWsgQfA2K6x5TLBUhI-1759889208-1.0.1.1-s9QE2J5GM086LUqeepDNyRorxAmC5.zm8zswVscY9Cd4uAhh_htzqrTRV0HZQdE6G2Cdy7hHpT7Z2Dm5LL37Rxkr6C2RC7sp3VQXTKxQO5Q; path=/; expires=Wed, 08-Oct-25 02:36:48 GMT; domain=.w3.org; HttpOnly; Secure; SameSite=None server: cloudflare alt-svc: h3=":443"; ma=86400 13:45:28 RRSAgent has joined #rqtf 13:45:33 logging to https://www.w3.org/2024/11/13-rqtf-irc 13:45:33 RRSAgent, make logs Public 13:45:34 please title this meeting ("meeting: ..."), jasonjgw 13:45:47 meeting: RQTF meeting 13:45:51 present+ 13:45:53 scribe+ 13:45:55 chair: jasonjgw 13:46:02 agenda+ Accessibility of machine learning and generative AI. 13:46:02 agenda+ Collaboration Tools Accessibility User Requirements. 13:46:02 agenda+ Miscellaneous topics. 13:58:31 Joshue108 has joined #rqtf 13:59:50 present+ 13:59:52 scribe+ 14:00:42 present+ 14:01:55 stacey has joined #RQTF 14:02:13 present+ 14:03:34 zakim, next item 14:03:34 agendum 1 -- Accessibility of machine learning and generative AI. -- taken up [from jasonjgw] 14:03:59 scribe+ 14:07:42 jgw: Introduces Josh ... 14:08:06 Joshue108: Can intro broadly 14:08:24 Joshue108: Finding scoping issues; much seems to be repair; believe we need goals beyond 14:08:37 Joshue108: Even blue sky, what the future could look like 14:08:59 Joshue108: Is that all there is?? 14:09:06 q+ 14:10:17 jasonjgw: Notes scope recently updated? Would your reaction change based on the update? 14:11:17 Joshue108 has joined #rqtf 14:12:12 Janina: scoping - had some creative ideas for future and being aware/helpful. Person with a profile, tracking, and directly helpful as a result of that. Notice certain features weren't invoked becuase use wasn't aware, offering mini courses. One possible way AI can do things that aren't easily achieved otherwise. Hoped to address some thoughts like 14:12:12 that. Open up TOC to leave room for things like that. maybe we're too subtle in our scoping. 14:12:58 Joshue108: other bits i'm intrigued by. That's what we should be looking at those things 14:14:01 Janina Discussed scoping, and we want to cover those topics. nice it can caption and describe (yes problems with that, less capture that), but it's the less interesting stuff. Interesting in what's possible, if we throw AI at it, what does it open up? 14:14:12 q? 14:14:15 ack ja 14:14:49 janina: Notes wider scope is intent; AI augmenting user skills being mindful of individual user's specific a11y needs 14:15:01 Joshua AI as a tool, not as it's there as fait accompli, but as an entity that's complete, it's really looking at what it can do, comes around to brute forrce processing, algorithmic potential 14:15:51 jasonjgw: Was looking whether GPT could provide info on how to accomplish tasks both cli or web 14:16:14 jasonjgw: Can answer linux cli questions; sometimes not accurate; sometimes helpful 14:16:28 jasonjgw: Sometimes provides answers more complex than could be 14:17:02 jasonjgw: tendency to generate inaccurate info is a concern; would be an elevated concern for COGA 14:17:50 Joshue108: will need to update my comments; phps obe? 14:18:04 Joshue108: Agree to expanding TOC to encompass wider opportunities 14:21:34 jasonjgw: Also q on where to fit ml models; obviously author side to help correct what's being said; all for the better 14:21:53 +q 14:22:03 jasonjgw: What we worry about is auto generated content without human qos 14:22:03 ack me 14:22:04 ack jo 14:22:38 Joshue108: Q: should address, distinctions between ml and generative 14:23:20 Janina: we have to be clear on what that means. I'm not clear right now, but need to be clear for the document 14:23:48 Joshue108: to be taken seriously we need to be clear 14:27:38 jasonjgw: One technique for developing AI is to use ML applying large lang learning modules 14:27:41 Jason: the way i think this is explained in the intro text, one technique for developing AI is to use machine learning especially neural networks (most widely used today) and one application is large language models, going beyond language, basic strategy is to predict what the next phrase/word is going to be in a sequence (prompts) 14:27:49 rajakushalnagar has joined #rqtf 14:28:27 jasonjgw: Based on large artificial large models 14:28:32 s/to be taken seriously we need to be clear/to be taken seriously we need to be clear about the distinction between AI and ML and what the strengths and weakness of each are in terms of their potential for accessibility 14:30:32 q+ 14:31:20 ack me 14:34:09 Janina: one side is machine learning (the benefit of GPUs, how it is achieved in the machine). It's an inference approach vs heuristic rule-based model? 14:34:15 q+ 14:34:45 +1 to Jason 14:35:01 jasonjgw: replacing how a neural system works with a probability based calculations that can be optimized to solve specific problems. Inferences run on them once created. 14:35:35 ack me 14:36:27 NLS Book: "Why Machines Learn: The Elegant Math Behind AI" 14:36:50 Joshue108: skirting around same thing. Generating things from stuff. Whatever info is there, AI, ML things will brute force process. CAn they anticipate user need? Change assitive tech? Hcnage support depending on context of use? How AI can be used to generate user interfaces and support users who don't have access to their assistive tech. 14:37:36 Joshue108: appears to be intelligent but brute force calculations 14:37:44 markovian chain 14:38:15 rajakushalnagar: Probabilistic calculations are part of the reason why AI can hallucinate. There hasn’t been a good way to reduce this other than to merge results from different models 14:38:50 rajakushalnagar: we have the probability, sometimes it guesses wrong and it hallucinates. And ongoing problem especially for a11y, bc you get the wrong results. 14:39:21 s/appears to be intelligent but brute force calculations/appears to be intelligent but really its just brute force calculations with some probability Markovian generated options sprinkled on top 14:39:24 rajakushalnagar: humans can also misunderstand and may assume incorrectly. But humans have more context typically. It's easier for humans not to guess incorrectly. With Ai, we don't know what the solution will be yet 14:40:18 rajakushalnagar: what's the best guess between two of them, less risk if fewer assumptions. Wrong guess for a particular model. Might be multiple models, risk of reducing...different models have to come to some sort of agreement. It's the guess between them 14:40:18 right, interesting Raja 14:40:45 raja: Reducing risk by recourse to multiple models 14:41:10 janina: Reminds of another book I need to understand: "The Alignment Problem" 14:41:24 q+ to say that especially in context of what Raja says it is important to keep some control so the users need is met, not the need of the AI 14:41:24 jasonjgw: ways of using them to correct each other to some degree (i think). techniques worked on recently engaging in a reasoning process. I don't know the mathmatics of neural networks. Could be interesting in p=improving accuracy. Area at forefront of reseach investment 14:41:24 janina: Has to do with optimization 14:42:04 raja: When models know everything about the world, results will be better; but not possible now 14:42:06 s/p=improving/improving 14:42:21 rajakushalnagar: fundamentally, models don't get good enough world data from print. Still missing the world contect because the model includes everything as it happens, but not possible right now. People think about what's important in the picture and speech. Interaction betweent eh two in the training is missing 14:42:26 ...what's the most important? 14:42:41 ack me 14:42:41 Joshue, you wanted to say that especially in context of what Raja says it is important to keep some control so the users need is met, not the need of the AI 14:43:03 Joshue108: We need to talk about what serves people's needs, not the needs of the tech 14:43:16 Joshue108: It's novel tech, and some would want to focus on that 14:43:27 Joshue108: Our view is what benefits folks, not what's cool tech 14:44:49 jasonjgw: Persistent worry then task of evaluating system output lands on pwd, but they're least able to do so 14:45:03 q+ 14:45:05 +1 to Jason 14:45:26 ack jan 14:46:54 Janina: how can AI help build smarter AT? We can generalize what constitutes better AT so we don't have to rely on the person with the disablity to say what's good. Every screen seems to be loaded...here expand this, look at this, but I want to squeeze in more chrome (scaffolding around, not the browser). All the stuff the provider is happy to do 14:46:55 for me. This is the bump in change your attention. 14:47:11 Stacey: I hate this purposeful friction! 14:47:33 Stacey: I get annoyed at designers for this 14:49:11 [discussion of using AI to declutter an overly busy, and tangentially related content] 14:49:38 Janina: items of scoping we want to discuss 14:50:35 zakim, next item 14:50:35 agendum 2 -- Collaboration Tools Accessibility User Requirements. -- taken up [from jasonjgw] 14:51:41 zakim, next item 14:51:41 agendum 3 -- Miscellaneous topics. -- taken up [from jasonjgw] 14:53:30 zakim, end meeting 14:53:30 As of this point the attendees have been jasonjgw, janina, Joshue, stacey 14:53:32 RRSAgent, please draft minutes 14:53:33 I have made the request to generate https://www.w3.org/2024/11/13-rqtf-minutes.html Zakim 14:53:42 I am happy to have been of service, jasonjgw; please remember to excuse RRSAgent. Goodbye 14:53:42 Zakim has left #rqtf 14:54:10 janina has left #rqtf 15:02:55 kirkwood has joined #rqtf 15:16:17 kirkwood has joined #rqtf