ALLi News Editor, Dan Holloway
I may be hallucinating this (see what I did?) but I think I’m right in saying that I didn’t report on a single AI-related story last week on ALLi's News Column. After a much needed lie down and a cup of tea, normal service is resumed. I will attempt to keep the AI round up to this one item.
Last week was London Book Fair, and it was inevitable that AI should feature prominently. In the end the prominence was slightly less than you might expect. Which is possibly an indication that its presence in our midst has normalised. Or possibly an indication of a new (downward) phase beginning in the hype cycle. The Guardian’s summary of themes at the Fair places AI third on the list, and focuses on the widespread discussion of the use of generative platforms like ChatGPT as assistive tools in the writing process. Which does indeed suggest we are reaching the early stages of a maturing relationship.
But one area where there was genuine excitement was around the European Union’s AI Act. In one of those “choreographed by fate” moments, the AI Act received its final approval on the middle day of the Fair. The Act has been widely welcomed by the publishing industry for pushing back many of big tech’s claims that self-regulation should be sufficient to protect everyone’s interests. At its heart is an insistence of transparency for generative platforms that means authors and publishers will know whether or not their work has been used in training AI.
A final interesting piece calls into question the progress AI is making in one of the areas that has so far garnered little attention in the cultural sphere despite being one of the earliest use cases for AI: translation. The European Council of Literary Translators recently recommended that translators avoid translating AI-generated text among other calls. One of the other statements emphasises that translation is a human art, and urges that AI translation falls short of that humanity. I am, as you know, a firm believer in solidarity among the creative corpus. I also believe claims about what “AI will never accomplish” tend not to age well.
