📝 Summary Report: Reflections on AI Companies and Sign Language Technology - Deaf Users’ Experience
- Tim Scannell
- 3 days ago
- 2 min read
After recent meetings with AI companies developing sign language technology, here are key reflections and calls to action shared by Deaf users and communities:
1. Introduction & Engagement
Too often, initial contact lacks transparency and accessibility. Deaf users need to understand:
Who you are
What is your mission?
How do you intend to collaborate (e.g., remote vs. face-to-face)
2. Commitment & Vision
Waiting for the “right moment” leads to delays. We need direction now:
Which sign languages are you supporting?
How are Deaf people actively involved?
What ethical frameworks (e.g., ISO, AI ethics standards) are guiding your work?
3. Deaf User Experience
From Sign Language Bibles to modern AI, this is a multi-generational journey. AI should support, not replace, human interpretation, and respect Deaf communication preferences and cultural norms.
4. Accuracy & Improvement
Just like Google Translate and YouTube captions, early versions can be flawed, but must improve:
Prioritise cultural/contextual accuracy
Use correct glosses and grammar
Ensure Deaf-led input, design, and evaluation
5. Access & Presentation
Whether using avatars, picture-in-picture (PIP), or autoplay features, consistency and user control are critical. Visual design must promote comfort and true accessibility.
6. AI Challenges
Ambitious goals include:
Real-time, bi-directional translation
Cross-platform integration (web, AR, HCI), but cost, quality, and usability gaps remain. Collaboration is key.
7. Language & Linguistics
BSL ≠ English. Sign languages are rich, visual, spatial, and deeply cultural. AI must understand full linguistic structures, not just hand movements.
8. Product vs. Tech Readiness
Some claim the “technology is ready,” but the product isn’t. One company is close to a working sign-to-text tool — but real-world testing with Deaf users is the next vital step.
9. Inclusion & Leadership
True inclusion means:
Hiring Deaf translators, engineers, UX testers
Co-designing with Deaf professionals
Shared decision-making, not token involvement
10. Standardisation & Approval
AI must:
Respect sign language variation
Follow Deaf-led approval processes
Rushing releases causes harm, not innovation.
11. Responsible, Ethical AI
Frequent issues include:
Misinterpretation
Cultural inaccuracies
“Hearsplaining” (hearing people explaining Deaf issues)
Let Deaf language and community needs lead the design, not hearing-centric norms.
12. Feedback & Transparency
Deleting critical comments ≠ unethical practice. That’s signwashing. Accountability means:
Listening to feedback
Being transparent about limitations
Building trust through honesty
13. Sustainability & Consequences
Ignoring Deaf input leads to:
Broken trust
Low user engagement
Accusations of tokenism (e.g., Disability Confidence levels 1–3)
Genuine engagement leads to:
Long-term impact
Community trust
Better design, better business
📢 Final Question to AI Companies:
Are you truly including Deaf people, or just inviting us after the key decisions have already been made?
🔗 Let’s build AI that listens, learns, and leads with Deaf voices at the centre.