Who knows where the audio track goes?

On a recent trip to the University of Leeds for the EC-TEL conference, I spotted an interesting sign on a noticeboard.

Image taken by me at University of Leeds. No rights reserved by me.

I looked up the Typeology website and read the following description:

Who are we ?
A team of members run Typology together as a co-operative. We are all graduates with a variety of qualifications and experience of academic environments. We are trained and experienced in audio transcription.

What is a co-operative.
A co-operative is an ethical form of business which is owned and run directly by its members. All our members have an equal say in running the business together. As we don’t have to answer to shareholders we can pay ourselves a fair wage while offering a competitive price to our customers.

A couple of things jumped out at me – that these were graduates who were well versed in the kind of content they might be asked to transcribe; and that as a coop they pay themselves a fair wage and can dictate the direction of their business.

Experience with existing human-mediated transcription services are that even where you pay for so-called 100% accurate transcriptions, often work is still needed. Specialist terminology or jargon unique to the HE environment often needs finessing. Having people well versed with that language (like students or recent graduates) would seem to be a thing well worth paying for.

Colleagues, including some of my team, have been running a small pilot to test various subtitling approaches and it occurred to me that I actually have no idea about the human mediated transcription that many of the big transcription suppliers provide (the kinds of suppliers that plugin to other systems and offer automated and/or human mediated transcription). Critically, I don’t know where this work is carried out, by whom, and whether they are paid a fair wage. I feel a bit ashamed after seeing this poster that I never really stopped to think to be honest.

As we push for greater use of subtitling for media in our Universities to improve accessibility for our communities, I’m worried that we might we be unwittingly be supporting precarious, poorly paid labour somewhere else. As services like AWS Transcribe start to be used more heavily, how big a leap is it to think that Amazon MTurk micro-workers might end up underpinning a human “value-add” service to improve transcripts? Checking out this kind of information suggests that’s pretty much a certainty, and some chat on Twitter also verifies that MTurk workers already pick up a lot of transcription work.

Colleagues kindly sent me on a bunch of reading to do in response to some tweets. So now I need to dig in and learn more.

Leave a Reply

Your email address will not be published. Required fields are marked *