Meeting minutes
<fabien_gandon> Tentative Agenda
<fabien_gandon> (1) Round-table: self-introduction of 30 seconds max for each participant.
<fabien_gandon> (2) Introduction to the IG by Roy (a 10 minutes).
<fabien_gandon> (3) Discussion and straw poll about the time slot and frequency of the IG meeting.
<fabien_gandon> (4) Open discussion and brainstorm session to identify topics and documents that should be on the radar of the radar IG.
Roundtable
gendler: welcome to the first Web & AI IG meeting, very excited to have you all here
… we'll start with a roundtable to introduce each other through 30s intros
… My name is Max Gendler, I'm the AC Rep for News Corp and one of the co-chairs of the IG
… I've been involved in various internet standards org
… AI is a topic of major interest for us as publishers
Fabien: I've been Inria's AC Rep since 2012, involved in W3C since 2004
… coming from a mainly academic backgroudn - senior researcher at Inria with a focus on knowledge graph and hybrid AI
… also interested in societal impact of AI, including attention capture
… also involved in Probabl, a company involved in providing open source tools for data science
Liang: I've been working in browser space in ByteDance for 5 years, before at Baidu
… involved in grpahics and performance optimization in Chromium browsers
… more recently focused on agent driven and browser automation
… how agents can efficiently operate on the Web
… interested in finding ways to better support agents on the Web
SteveF: co-founder/director at Tetralogical along with Léonie and Henny
… involved in Web standards since 2007
… incl HTML5 spec, and accessibilited-related browser specs
… particularly interested in accessibility testing and how/whether AI can be used for that
… looking for use cases around accessibility testing
Roy: work for W3C based in China; staff contact for the IG
… I'm pursuing my PhD degree on Web & AI, and accessibility for computer science
Andrei: assistant professor in Switzerland and also involved in Wimmics at Inria in France
… research focused on Web agents
… co-chairing the Web Agents CG
… also following Web of Things work in W3C, and involved in IETF as well
Léonie: work for Tetralogical; interest from 3 different angles: as a blind user, very interested in how AI has changed and is changing the way I use the Web
… interested in what standards might be needed for AI on the Web: e.g. conversational interfaces for the Web
… finally, impact of AI on us Tetralogical as a company running accessibility audits - moving away from a well-defined set of pages with well-defined content
AlexanderD: sustainability advocate, editor of the Web Sustainibility guidelines in the related IG
… also involved in Green Software foundation
… in HTTP Archive
… interest primarily around holistic impact of AI on environment, people (societal & governance)
DanielD: joining from WU Vienna (Austria)
… interested in rediscovering the original vision of the Semantic Web with integration of AI Agents
Vagner: working as a senior researcher at NIC.br; in Machine Leaning and responsible AI
cpn: work at BBC, also co-chair of the MEdia & Entertainment IG at W3C, W3C Media WG
… BBC is in an interesting position, with its content beind used and intermediated by AI systems which concerns us in terms of access to reliable and accurate journalism
… but we're also making use of AI in our internal workflows
… standardization-wise, I'm active in related work in IETF: AIPrefs and WebBotAuthn
… both looking at the relationship between web sites and those that crawl these web sites for AI
Henny: also at Tetralogical
… have contributed to W3C years ago in WAI EO & UAAG, really interested in relationship between AI & accessibility
… how it can enable and disable people
… and impact on our day to day work on accessibility assessments
… how our customers might be using AI to build their web sites
Rem: associate profession at University College in Dublin, involved in multi agent systems since the 90's
… trying to combine traditional multi agent systems with agents/LLms
… also co-chairs Web Agents CG with Andrei
Paola: co-chair of the AI KR CG
… [inaudible]
Craig: work at Tetralogical as a Principal accessibility specialist; previously involved in UK gov, tech companies (ElasticSearch as they introduced AI)
… interested in pulling apart the promises of AI and how much they are met through lots of hands-on experimentation
… I use it a lot, but also have a lot of concerns on the hype around it and see how standards can help
Ahmet: I'm with FIME, which provides consultanty services in payments
… mostly interested in how agentic solutions communicate with the Web
… we've been looking at solutions based on MCP servers
… interested in how this will evolve in the future; how and whether human-intended interactions evolve to accommodate agents
Mihai: researcher at DFKI in Germany, focused on responsible and ethical development of AI systems
… I contribute with different institutions and standard organizations on ethics related initiatives
… incl BDVA and European Executive Agency
… very intriguted by the joint deliverable on Ethical Principles for Web Machine Learning - looking forward to finding more about it
Dom: I'm Dom, member of the w3c team. I've been looking over impact of AI on the web ...
started with a workshop and a report on the AI systems impacts on the web in 2024...
… more recently, impacts of agents on web. Working with Roy and other in the community to get the group launched.
Ehsan: from Samsung R&D with a focus on privacy
… also editor of the Translation API and a TAG associate member to review APIs
… I have an academic background - an assistant professor at Durham University on privacy-by design, socio-technical trust,
… interested in privacy discussions
Shouqun: AC Rep of Bytedance - we're involved in Immersive Web, WebRTC
… engineer lead of the cross-Platform team based on an open source framework mostly based on Web technologies
… we're interested in generative UIs
Diogo: I work for NIC.br (also host the W3C Brazilian Chapter)
… interested in the impact of AI on the Web on governance, also economic and technical impact
Round-table
<Roy_Ruoxi> Web & AI IG Overview slides
Introduction to the IG by Roy
Slideset: https://
Roy: the IG was started as a way to support the many AI-related conversations we saw happen in W3C
Roy: some groups have tried using AI for their meeting mintues - maybe something we can investigate as well
Roy: our work will be in public, so also available to non-participants
Roy: Dom & I have been looking at a workshop on the topic of "AI Agents & Web browsers", still work in progress
… we're looking for more input, incl a programme committee - we hope this IG can help
Roy: I've been working on a draft Web & AI roadmap that I hope to share with the IG soon
Roy: as staff contact for this IG, please feel free to contact me if you have any question
Discussion and straw poll about the time slot and frequency of the IG meeting.
Gendler: we're thinking of an initial monthly cadence for this meeting
… would Monday around (or exactly at) this time work for you? as a temperature check, please +1/-1 on IRC
<Rem> +1
<cpn> +1
<aciortea> +1
<fabien_gandon> +1
<Roy_Ruoxi> +1
<Craig> +1
<Henny> +1
<gendler> Straw Poll: Meeting on a monthly cadence, on Mondays, at around the time of the first meeting
<dom> +1
<stevef> -1
Gendler: we will also get feedback from people who aren't here today
<santanavagner> +1
Steve: this conflicts with an AGWG subgroup meeting
Open discussion and brainstorm
Fabien: what topics would you look to see discussed in this IG? what deliverables/reports would you like to help create?
… this is a brainstorm to collect ideas, not discuss them in depth
… please q+ to mention the topics you have in mind
Dom: more on the logistics, not that I don't have topics...
… several people mentioned topics in the intro, I hope we bring that into the list.
Rem: MCP, and intersection of agents/LLM with the Web feels like an interesting topic; we have had related discussion in the Web & Agents CG
Dom: one of the topics people ask me about, is the infrastructure impact of crawlers on the web...
… more and more types of agents and LLMs are interacting with the web, and are creating business model challaenges and infrastructure challenges....
… it can reduce potential revenue, but also increase costs due to amount of crawling, both a W3C and IETF topic....
… it would be useful to see how it looks on traditional items of skope like APIs.
<Zakim> dom, you wanted to suggest impact of crawlers
<fabien_gandon> ack Roy_Ruoxi
Roy: how to set up bridges & liaisons with different organizations: IETF, Linux Foundation, since they have AI & AI agents related work there
… we need to find the right persons there, and the relevant topics for the Web
<cpn> Re Dom's topic, here's the IETF discussion on crawler impacts: https://
Ahmet: MCP was mentioned - there is a growing interest in agentic e-commercce, with shopping agents / shopping assistants
… UCP from Google was announced yesterday (Universal Commerce Protocol) to enable merchants to talk with agents
… interested to understand impact of adoption of these types of protocols
… in addition to CMp
<Zakim> fabien_gandon, you wanted to mention identifying the impact AI at every level of the Web architecture client, server, proxy, application, service, infrastructure and devices, etc.
fabien_gandon: among the topics, all the topics on the Web as a space to host interactions with these agents
… human-to-agents, agent-to-agents, agent-to-resources
… impact on infrastructure & its cost
… also strong interest around traceability of AI applications & content, with growing demand incl from people building models
… in terms of organizing the discussions, I would be happy if we had enough expertise to cover all aspects of AI on the architecture: client, server, proxy, apps, etc
DanielD: +1 to Fabien; also looking at existing standards in the agentic ecosystem
Craig: the rise of vibe coding and its impact on accessibility, HTML would also be useful to discuss
… with AI replacing standards user interface - what impact on WCAG?
<fabien_gandon> (we should close the queue)
Shouqun: Google released A2UI to generate user interfaces
… keeping track of these protocol evolutions would be useful
<cpn> +1 fabien on traceability, also content provenance technologies
AlexDawson: it would be great to have coverage on what happens when we don't have data, e.g. because of lack of transparency
… that has an impact on sustainability considerations for instance
<fabien_gandon> +1 to Alexander point also what AI will do to all types of divides (eg low-resource languages) and sustainability problems
AlexDawson: sustainability is less mature than accessibility in that respect
Max: thank you all, very excited to get this work rolling - please reach out for any additional topic suggestion