Recently, I had to prove that I’d written 300,000 words myself. Why? I’m now a qualified member of the UK Institute of Translation and Interpreting (MITI) for all my three source languages: Finnish, German and Polish into English. If you translate into or out of the “big four” languages, or FIGS (French, Italian, German and Spanish), you qualify as a member by doing an exam. The assessment is as close as possible to a real translation assignment. You get a text, a couple of days to research and translate it, and then you submit it, with a commentary explaining your decisions. I did this for German into English years ago, and now I assess other translators’ assignments. But if your working language is not so easy to study in the UK, there are not enough assessors to mark the exam. Helping with assessment was my motivation to qualify as a MITI in two “smaller” languages (Finnish and Polish). That process was different: I had to gather a portfolio of evidence about my translation work. But if I say I’ve translated something, how can they know it was me? It should be easy enough, when translators are named. While my name is in the published books I’ve translated, for other texts, we don’t always know the translator. The author or publisher of that writing might not acknowledge their work. Or the translator might not be able to say, if the document is confidential.
Proving it was you is getting difficult for other kinds of writers. A few weeks ago Lizzie Wolkovich explained in Nature how a journal wrongly accused her of using ChatGPT to write a research paper. It works the other way too; large language models are using authors’ work without their consent, breaching copyright. In its response to this issue, the UK Society of Authors is clear that “machines cannot be authors. Our copyright regime relies on concepts of human originality and skill and labour. Only humans can create and receive copyright protection.” The Russell Group universities in the UK have come up with very broad principles on the use of generative AI in higher education. They raise ethical concerns and make a commitment “to ensure that academic rigour and integrity is upheld.” In such a fast-changing field, the practical policy is much harder work.
If you write yourself, assess people’s writing, or help others to write better, I’m sure you’ve been thinking about this. I have, too, with my editor and translator colleagues, and the academics I work with. We’ve been talking about how machines affect human writing at writing retreats, on training courses, and in committee meetings. While some have already had more than enough of it, the conversation is only beginning. Shaping policy is hard work – but exciting. We can prove the value of human originality and skill and labour. When you write, you should be able to prove it.
Ridge Writing Retreats 