16 January 2015

Duolingo: Web 2.0, free labour and the power of ignorance

Last time I wrote anything here, I had decided I was going to get some German under my belt. So I've tried out a couple of things on the net, and I've spent a lot of time on Duolingo, which in many ways is a very good resource, but is frustrating in the way it keeps generating nonsensical phrases and fragments.

Well, it turns out they recently added an interesting clause to their user agreement:
Temporary Restrictions on Users from the European Union
Users within the European Union are not presently allowed to submit materials for translation or translated materials to Duolingo. While these users can continue to use the educational services offered through the Website, they will not be involved in the translation of any documents. If you submit a request for translation or translated materials to Duolingo, you thereby warrant and represent that you are not currently within the European Union, did not translate the document within the European Union, and will not be within the European Union when your translation request has been finalized.
So what's going on here then?

I have always felt that most dot-com organisations run on a model that breaches workers' rights laws. In most countries, a for-profit organisation is not allowed to solicit or accept free labour, and yet a great many commercial internet sites rely on free labour for their profit.

When YouTube first launched, all advertising revenue was kept by the site -- uploaders made no money. YouTube argued that the uploaders weren't working for the site, so didn't need paid... and yet, without the uploaders, there would be no site. YouTube changed their business model later on to grant uploaders a share of the advertising take. The reason they did this was so that they could get on board professional media (including music videos) and then also to stop the higher quality amateurs from migrating to sites that were willing to split the profits. Market forces worked in the interests of the little guy... this time.

But what about Facebook's big translation push at the time of the public share offering? The public sale brought in enough money to translate the site into all the world's major languages several times over, and yet they did not pay a single translator, instead "crowdsourcing" the translation. It would be one thing if they had opened translation to any and all languages, but they chose the languages and were only interested in the "big" languages that would draw plenty of users and make Facebook more money.

If it was small languages, I could understand: you don't want to pay for a translation to eg Irish when all the users will happily use the English version -- it doesn't make you any money. But when you're translating into Spanish, one of the world's most widely spoken native languages, you'll make your money back many times over even if you pay for one of the world's best translators.

Facebook clearly thought that if so many other websites had got away with free labour, they would too, but they inadvertently brought the issue to far more public attention than they expected.

You see, translators have real power in Europe. With such a linguistically diverse base, the institutions of the European Union are full of translators, which makes them one of the most powerful lobby groups you can imagine. Seriously, there is no-one who "has the ear" of a Brussels bureaucrat than the person who's talking in that ear throughout the meetings.

Now I don't recall ever hearing of any sanctions being made against Facebook for this, but the groundwork was set and Duolingo walked right into the problem, because more than any other site, their business model is built on unpaid labour... and crucially unpaid translation. Duolingo seeks to generate income by having learners translate documents for paying clients as part of their "immersion" in the language. Already, Duolingo is translating articles for Buzzfeed and CNN. Their justification is that the translators are getting something in return -- the teaching. I can see where they're coming from, to a point, but that's the same justification people try to make for internships as a source of unpaid labour.

So somewhere along the line, Duolingo has been warned off and put up these "temporary restrictions"... but didn't tell anyone about it. It's there, right at the end of the Ts&Cs, but they didn't actively notify users, and there is no notice on the translation page to warn you that you might be about to do something potentially illegal.

But it gets worse, because they don't only leave you access to the section that is illegal, but they actively encourage you to use it. I've been using it a lot recently, and after most exercises, it tells me to try the translation.

Now, if you're sitting at a computer in the UK and try to access BBC Worldwide clips on YouTube, you won't get anything. Why? BBC Worldwide content is licensed for use outside the UK, and YouTube knows where you are. The same thing happens on plenty of sites.

Duolingo makes no attempt to block based on location, but there is no technical reason that they shouldn't. I cannot imagine that a company their size would not be tracking user locations anyway, in order to optimise their marketing strategies and their technology. They must know. Furthermore, there is even a section in the profile (optional, admittedly) for you to tell them where you live.

It's a pretty stupid course of action, if you ask me. With geolocation being such a simple and standard admin task (although admittedly not 100% accurate), failure to attempt to identify and block EU-based users could be argued to be negligent. That negligence is surely made worse by the fact that they are leading their users not only to arguably (not tested in court) break the law, but also to indisputably break their own license agreement. And all the while their negligence allows them to continue selling translations to commercial clients.

It's a dangerous path, and it could lead to a very messy end....

No comments: