This year’s winner of the respected literary magazine Granta’s competition for unpublished short stories from the British Commonwealth was Jamir Nazir’s “Serpent in the Grove.” The panel of judges, headed by award-winning writer Louise Doughty, chose his entry from among 7,806 others. Then along came a professor studying the impact of artificial intelligence on the labor market and education, who ran the story through Pangram, a program that can spot AI-generated content with 99% accuracy, and came to the conclusion that it had not been written by Nazir at all, but by AI.
This incident illustrates how difficult it has become to distinguish human writing from that produced through algorithmic repetition. Literary critics went so far as to describe it as a “Rubicon moment.” We read of such incidents every day. A book that was recently published in the United States on the effect of AI on truth, titled “The Future of Truth,” was found to contain quotes generated by AI. Writer Steven Rosenbaum’s admission that this is the best possible confirmation of the risks he raises in his book seems a lot like justification after the fact.
The New York Times, meanwhile, severed ties with a literary critic after discovering that he used AI to write a book review, so that several parts ended up being identical to those of a colleague at The Guardian. These incidents point to a perpetual battle to define the relationship between the media and AI, which can either constitute a valuable tool or a tombstone for the media. The relationship between politics and AI may prove equally dangerous. US President Donald Trump appears to be the only major world leader who has embraced the potential of artificial intelligence with such uncritical vigor. He is, after all, the politician who introduced the world to the “post-truth” of 2016, so he can be said to be consistent in his principles.
