MAP Workshop
Workshop

Multimodal Accessibility Planning with AI Models

A focused workshop around multimodal AI evaluation, accessibility-aware planning, and evidence use for real-world place information.

Workshop Information

We provide details on the public workshop for the benchmark and challenge.

Click here to register for the workshop.

Important dates

Workshop date
October 25, 2026.
Application deadline
Thursday, August 27, 2026, 11:59 p.m. Anywhere on Earth (AoE, UTC-12:00).
Notification date
Thursday, September 3, 2026.
Registration
All accepted participants must register for the full ASSETS 2026 conference and add the workshop during registration.

Schedule

Note: times below include am and pm slots (pending confirmation on whether the workshop will be in the morning or afternoon)

  • 9:00-9:30 / 13:00-13:30
    Welcome address, presentation of the challenge and challenge results.
  • 9:30-10:00 / 13:30-14:00
    Keynote: Sasha Blair-Goldensohn, Google Maps Staff Accessibility Lead.
  • 10:00-10:30 / 14:00-14:30
    Keynote: Ross Greer, University of California, Merced: "Automated Context-Aware Navigation Support for Individuals with Visual Impairment Using Multimodal Language Models in Urban Environments".
  • 10:30-11:00 / 14:30-15:00
    Coffee break.
  • 11:00-11:30 / 15:00-15:30
    Keynote: Jon Froehlich, University of Washington: "Vision Foundation Models and Generative AI for Accessibility: Challenges and Opportunities".
  • 11:30-12:00 / 15:30-16:00
    Panel discussion:
    • Robin Angelini, PreDoc Researcher, TU Wien
    • Jessica Rafuse, Accessibility Leader, Microsoft
    • Sarah Ebling, Professor of Language, Technology and Accessibility, University of Zurich
    • N.N.
  • 12:00-13:00 / 16:00-17:00
    Break-out groups: planning the 2027 edition of the challenge.

Diversity and Inclusion Considerations

We are aiming for diversity in gender, disability, and other protected characteristics as well as geographical background (specifically looking to welcome participants beyond the US and Europe) at the workshop, both among the contributors (keynote speakers and panelists) and the participants.

Organisers

The organising team consists of a mix of representatives from academia and industry:

  • Jason Armitage, University of Zurich, Switzerland
  • Ioannis Tsochantaridis, Google DeepMind
  • Zhiyong Zhou, Zhejiang University, China
  • Linda Mazzone, University of Zurich, Switzerland
  • Chuqiao Yan, University of Zurich, Switzerland
  • Srini Narayanan, Google DeepMind
  • Sarah Ebling, University of Zurich, Switzerland

Call for Participation

We invite participation in the ASSETS 2026 "Multimodal Accessibility Planning with AI Models" (MAP) Workshop to be held on October 25, 2026 in Porto, Portugal. The workshop is linked to a challenge funded by Google.org on multimodal AI models - specifically, Large Language Models (LLMs) with vision input support such as multimodal versions of GPT, Gemini, or Qwen - assisting persons with disabilities in planning an accessible activity in the real world, such as finding a doctor, a museum, or a supermarket that fulfill certain accessibility requirements. The aim of the challenge is to assess the safety and usefulness of model responses. The challenge runs in July/August 2026.

The workshop will consist of a presentation of the challenge results via posters as well as three keynote talks and a panel discussion as well as a break-out session aimed at planning the next iteration of the challenge in a community-driven manner. Preference is given to participants with demonstrated prior exposure to the topic and potential to contribute to the diversity of the workshop. We will seek to publish the motivation statements on the workshop website prior to the event. We explicitly welcome participants from both academia and industry/accessibility practice and from all over the world.

How to apply

To apply for the workshop, please submit a motivation statement by email to Sarah Ebling at ebling@cl.uzh.ch.

The motivation statement should be either:

  • 2-4 pages of text;
  • an audio recording with spoken equivalent content; or
  • a signed video of 4-16 minutes in International Sign or American Sign Language.

Applications must be received by Thursday, August 27, 2026, 11:59 p.m. Anywhere on Earth (AoE, UTC-12:00). We expect to notify applicants by Thursday, September 3, 2026.

Access requirements

For questions about access requirements for participating in the workshop, please contact Jason Armitage at jason.armitage@uzh.ch.

Please contact us as early as possible so that we can plan appropriate support.