12:32:37 <RRSAgent> RRSAgent has joined #rqtf
12:32:41 <RRSAgent> logging to https://www.w3.org/2025/07/02-rqtf-irc
12:32:41 <Zakim> RRSAgent, make logs Public
12:32:42 <Zakim> please title this meeting ("meeting: ..."), jasonjgw
12:32:48 <jasonjgw> meeting: RQTF meeting
12:33:03 <jasonjgw> chair: jasonjgw
12:33:05 <jasonjgw> present+
12:33:07 <jasonjgw> scribe+
12:33:29 <jasonjgw> agenda+ Accessibility of Machine Learning and Generative AI.
12:33:41 <jasonjgw> agenda+ Miscellaneous topics.
12:52:53 <jasonjgw> jasonjgw has joined #rqtf
12:53:14 <jasonjgw> q?
12:53:27 <jasonjgw> agenda?
12:56:21 <scott_h> scott_h has joined #rqtf
12:56:28 <jasonjgw> scribe+
13:00:33 <scott_h> present+
13:01:13 <JPaton> JPaton has joined #rqtf
13:01:36 <stacey> stacey has joined #rqtf
13:01:42 <stacey> present+
13:02:00 <JPaton> present+ John_Paton
13:02:28 <scott_h> scribe+
13:02:30 <janina> janina has joined #rqtf
13:02:33 <stacey> scribe+
13:02:36 <janina> present+
13:03:15 <jasonjgw> zakim, next item
13:03:15 <Zakim> agendum 1 -- Accessibility of Machine Learning and Generative AI. -- taken up [from jasonjgw]
13:03:42 <scott_h> jason: bibligrophy update
13:04:06 <scott_h> looking to create automatically generated bibliogrophy based on code has been successful
13:04:13 <scott_h> works as intended, not difficult to use
13:04:36 <scott_h> Uses JavaScript to load bibliography file
13:04:48 <scott_h> but needs to serve content over HTTP rather than local file
13:05:09 <scott_h> to test locally, create local web server
13:05:54 <scott_h> second point: in actual text, uses identifiers of records as citation text
13:06:25 <scott_h> issue with exported citation is labels unrelated to article
13:06:49 <scott_h> just need ot rename those and edit out records
13:06:59 <scott_h> otherwise it works and shoudl work with GitHub
13:07:05 <scott_h> woud like moe time to fine-tune
13:07:48 <scott_h> janina: any otehr way to do it locally?
13:08:41 <scott_h> jason: sourcing files needed - bibliogrpahy, JavaScript, CSS, indes.tml file that is our document
13:09:07 <scott_h> need to use local web server or JavaScript files
13:11:11 <scott_h> janina: so repoistory has these files?
13:11:13 <scott_h> jason: yes
13:11:29 <scott_h> janina: so all we need is bibliography text?
13:11:32 <scott_h> jason: yes
13:13:06 <scott_h> janina: for seocnd point, each record needs to be edited by hand?
13:13:32 <scott_h> jason: yes, human readable name needed
13:13:50 <scott_h> jason: limitation of the exporting tools in how they label fields
13:16:32 <scott_h> janina: what's left ot test?
13:16:50 <scott_h> jason: convert over records cited in teh draft and  commit to a branch
13:18:15 <scott_h> update 2
13:18:23 <scott_h> went to the wiki and added references and may add some more
13:19:25 <janina> Related section in Edge Capabilities: https://A11yEdge.github.io/capabilities/#intent
13:19:27 <scott_h> janina: did some work in capabilities, see link
13:19:47 <scott_h> first draft but relevant
13:20:17 <scott_h> websites being replaced with intent-driven design/user interfaces
13:21:22 <scott_h> task and user-population specific on how you build an interface
13:22:05 <scott_h> interface is built to maximise funcitonality for a target audience, e.g. accessibility, for partiuclar population
13:22:14 <scott_h> and works by leveraging APIs
13:29:01 <scott_h> stacey: key is not if AI is in a product, key is if it works to achieve what users want
13:29:33 <scott_h> on development side, AI is causing a loss in critial thought
13:30:36 <janina> example: https://innosearch.ai/
13:33:38 <scott_h> jason: having ML tools to help with development coud make this apporach more achievable
13:34:32 <scott_h> literature suggests AI can have accessibility go beyond standards limitations
13:34:52 <scott_h> interesting and contriversal
13:36:35 <scott_h> e.g. if AI could digest a website and provide that content in an accessible way, could be accessibility solution
13:39:37 <scott_h> jason: related issue - machine learning-based agents
13:40:39 <scott_h> if agent has access to APIs
13:41:03 <scott_h> interacting with human user at one end, APIs at the other, traditional website may not be required
13:50:49 <scott_h> jason: do wonder what happens when hype goes awa and technology c continues to evolve
13:53:32 <scott_h> stacey: two different AI appraoches - AI to replace what humans do vs AI as a companion tool
13:54:52 <stacey> scott: doc started how can Ai improve in current tech, how can it support remediation in web content before user accesses it, now what happens if no website at all? straight from human need to API machine learning and interface of whatever the person wants.
13:55:42 <stacey> scott: if AI is effective maybe the website isn't required at all for accessibility? great discussion
13:56:06 <stacey> jason: interactivng with agent
13:56:15 <stacey> scott: no curated website
13:57:00 <stacey> jason: need to say something on agents, not sure where it is in current structure
13:57:12 <jasonjgw> zakim, end meeting
13:57:12 <Zakim> As of this point the attendees have been jasonjgw, scott_h, stacey, John_Paton, janina
13:57:14 <Zakim> RRSAgent, please draft minutes
13:57:15 <RRSAgent> I have made the request to generate https://www.w3.org/2025/07/02-rqtf-minutes.html Zakim
13:57:22 <Zakim> I am happy to have been of service, jasonjgw; please remember to excuse RRSAgent.  Goodbye
13:57:22 <Zakim> Zakim has left #rqtf
13:58:00 <janina> janina has left #rqtf