The events in Myanmar remind us that language access in technology is a human rights issue

The news from Myanmar and the failings of language access across the board remind us just how vital supporting language access in our technologies is to basic human rights. Here are some draft principles that might help us get there.(This is a follow-up that builds on a recent post noting that Translation on the internet is a full stack problem.)

In that post, I explored the idea that issues around hate speech and translation in Myanmar reflect the deeper truth that language barriers online reflect a compounding series of issues — from fonts to keyboards to content moderation — that exacerbate existing inequalities.

This past April at RightsCon, I organized a workshop and panel with Kate Coyer from Central European University, Dragana Kaurin from Localization Lab, Jovan Jelicic from MercyCorps and my colleagues Ed Bice and Tom Trewinnard from Meedan. It was a powerhouse group of some of the world’s leading thinkers on language, technology and rights, matched only by the brilliance of the attendees. It built off of panels we’ve done over the years at RightsCon examining basic principles of language access in technology. Language access is an all-too-overlooked matter of basic rights, and the stories coming out of Myanmar are a painful reminder of how underresourcing language support creates conditions for online hate speech to go unchecked.

In our workshop, we put together a few principles. They’re very rough, and we only had an hour, but I hope they can serve as a backbone for a more specific set of principles and standards that every technology developer takes seriously in a digital age where hate speech, misinformation and propaganda can flow in any language but basic support does not.

Some of the big themes?

• Context matters just as much, if not more so, than content.
• Language use online is multimodal: it can be text, audio, video, performance, emoji.
• Technologists should build with, not for, communities and understand local needs and contexts.
• Language access is a matter of power and power dynamics.

And here are the principles in raw form, transcribed directly from the workshops:

Group 1

• Language access efforts must respect context.
• Communications should be able to meet the needs of affected populations/recipients.
• Successful tools will enable multimodal expression and enhance trust.
• Successful communication requires trust, built on local recipients — conscious/aware, context-sensitive interpretation.
• Creative modes of communication may be the most effective; theater can be a more effective form of translation than precise [word-for-word] translation.
• Empathetic translation requires a deep understanding of context, for which we must be take the time to hear who we’re talking to.
• Effective translation requires trust, understanding, and connection between speaker and translator.

Group 2

• Language is a component of power. It is never neutral.
• Language access it not just content. It is also context.
• Language access in tools should be built with, not for, communities that speak it.
• Technology usage and language are interconnected.
• There is an endless array of problems. We have to be open to adaptability.

Group 3

• Language equity be closer to our reach when technologists and those with field experience work closer together.
• Language access in our technology is a human right because South to South communication strengthens movements, i.e., strategy, tactics and information sharing.
• When it comes to language access, content moderators should take into consideration cultural context.
• Language equity is about creating tools/processes that are iterative, flexible, and able to shift as needed.

Group 4

• When it comes to language access, designers should consider audiovisual elements of communications (including memes, color, arrows and visual emojis)
• Language equity will be achieved when solutions are created and designed for a need that exists in a country.
• Language equity can be achieved when technologists acknowledge power dynamics.
• Language access in our technology is a human right because without language there is no access.
• Language equity will be achieved when technologists don’t have to code in English.

These are critical issues. When I was doing logistics for events, I was tasked once with ensuring the lights and audio worked well. It was a thankless job—no one noticed when they worked well, but everyone complained when they didn’t. Translation and language access online are like the lights and audio of an event. They become obvious only in their absence.

As many in the rights community have noted, language access is a human right. Fernand de Varennes has argued at UNESCO:

Language rights are not collective rights, nor do they constitute ‘third generation’ or vague, unenforceable rights: by and large, the language rights of minorities are an integral part of well established, basic human rights widely recognised in international law, just as are the rights of women and children.

Technologists can and must do better, and we have to outline exactly what it means to make language access in tech a basic right. We’re far behind in that regard.