In a 1994 interview, the writer and sci-fi sage Ray Bradbury said that one of the motivations for writing his most famous book Fahrenheit 451, about a dystopian society where books are outlawed, was the scourge of “political correctness” which could lead to censoring of books. In previous interviews, when talking about the 1953 book, he had cited the paranoia of the McCarthy era and the effect of mass media on literature as having spurred him to write. Had Bradbury been alive today, one wonders if he would have perceived the danger posed to books by artificial intelligence (AI), and seen in it the potential for the execution of an age-old threat via a very 21st century means.
There is something to worry about in an Iowa school district’s solution when confronted with a new law passed by the state which requires school library catalogues to be “age appropriate” and contain no “descriptions or visual depictions of a sex act”. The authorities at Mason City Community School District put a “master list” of the most challenged books in the US through ChatGPT so that the programme could identify those that wouldn’t comply with the new law: 19 books — including Alice Walker’s The Colour Purple, Margaret Atwood’s The Handmaid’s Tale and Toni Morrison’s Beloved — have been removed from its libraries.
Since ChatGPT and other generative AI programmes started making headlines last year, several terabytes worth of arguments have been made both in the technology’s favour and against it. The pendulum has swung from fantasies about its potential for medicine, space technology etc to spectres about its impact on jobs and human creativity. But as the Iowa school district’s example shows, what is needed, really, is caution: When put in the hands of flawed humans, any tool – which, after all, is what AI is — can become a weapon.