🚨 ALERT: “Real-Time” Sign Language Claims Are Misleading — and Avoid Safer Options

There’s a growing push to use AI sign-language avatars in transport systems. But there’s a serious problem with how they’re being advertised:

❌ AI sign-language is NOT real-time

❌ It does NOT interpret live tannoy announcements

❌ It does NOT support urgent, live communication

❌ It is NOT equal access

So why use the phrase “real-time translation” when it isn’t live? We need to question this — clearly and loudly.

Image Description (Alt-text) A Deaf person stands alone among faceless crowds, surrounded by symbols of trains and sound. Their expression shows the stress of missing urgent information in a world built around audio. The contrast between the signer and the anonymous masses highlights that AI avatars cannot replace real, safe access to communication.

⚠️ There are safer, human-led options transport could use right now

If we truly care about accessibility, why aren’t we seeing:

✔ Instant interpreter access via kiosk QR codes

A Deaf passenger scans the code, connects to a live interpreter, and gets support within minutes — critical when a train leaves in 30 minutes.

✔ Customer service video booths with remote BSL support

Other sectors use remote interpreting every day. Why not transport?

✔ Support through mobile apps (Trainline, BSL platforms, in-house interpreters)

The infrastructure already exists. Real humans exist.Real interpreting exists.

So why pretend AI can do what it cannot?

🚉 AI must reflect real reality, not the marketing version

AI sign-language systems today can support:

✔ Pre-information – scheduled messages, timetables, planned disruptions

✔ Post-information – incident summaries, follow-up updates

But they CANNOT support:

❌ Live tannoy announcements

❌ Emergencies

❌ Last-minute platform changes

❌ Driver/guard instructions

This gap is not a small issue.This is a public safety issue.

🔐 Safeguarding is essential

If AI is used in stations or public safety environments, it must be:

✔ Accurate✔ Linguistically verified✔ Safe for emergency use✔ Clear about limitations✔ Accountable✔ NEVER marketed as “live” unless it truly is live

Using the phrase “real-time translation” is misleading and undermines trust.

📣 Questions transport operators & investors MUST ask

1️⃣ Does this translate live audio → BSL, or only text?

2️⃣ What happens when a tannoy announcement is made before text updates?

3️⃣ What safeguards protect Deaf passengers during disruption?

4️⃣ Why not offer instant access to human interpreters via kiosk or QR?

5️⃣ Who verified the language accuracy and safety of this system?

If these answers are unclear, the technology is not ready for real-world inclusion.

💬 I am not against AI — I am against misrepresentation.

AI has value.AI can support pre-written information, digital access, and consistency.

But AI is not a live interpreter.

Not today.

Honesty builds trust.

Exaggeration destroys it.

Accessibility must always be:

Human-led.Truth-based.Safe.Transparent.