Fable, a popular social media app that describes itself as a place for “bookworms and bingewatchers,” has created an AI-powered year-end summary feature that describes what book users are reading in 2024. It was supposed to be playful and fun, but some iterations took on a strangely combative tone. Briefing writer Danny Groves, for example, asked if he “remains in the stereotype of a straight, cis white man” after calling him “a person who is committed to diversity.”
Literary activist Tiana Trammell’s summary, meanwhile, concludes with the following advice: “Don’t forget to pop in on a white writer every once in a while, OKAY?”
Trammell was surprised, and soon realized he wasn’t alone after sharing his experience with Fable recaps on Threads. “I got a lot of messages,” she said, from people whose resume made inappropriate comments about “disability and gender orientation.”
Since the debut of Spotify Wrapped, year round features have become ubiquitous across the internet, providing users with lists of books and news articles they’ve read, songs they’ve listened to, and exercises they’ve completed. Some companies are now using AI to completely automate or enhance the way these metrics are presented. Spotify, for example, now offers an AI-generated podcast where robots analyze your listening history and make predictions about your life based on your preferences. Legend jumped on the bandwagon by using OpenAI’s API to generate summaries of the past 12 months of its users’ reading habits, but it didn’t expect the AI ​​model to spit out comments that took the anti-woke pundit seriously. .
Fable later apologized on multiple social media channels, including Threads and Instagram, where he posted a video of the officer issuing a mea culpa. “We deeply apologize for the inconvenience caused by some of our student briefs this week,” the company wrote in a statement. “We will do better.”
Kimberly Marsh Allee, Fable’s head of community, told WIRED that the company is working on a series of changes to improve its AI briefs, including an opt-out option and clear disclosures that indicate they are AI-generated. “Currently, we have removed the part of the model that engages the reader in play, and instead, the model simply summarizes the user’s taste in books,” he said.
For some users, fixing the AI ​​doesn’t sound like a sufficient answer. Fantasy and romance writer AR Kaufer was surprised to see screenshots of some of the sketches on social media. “They must say that they are completely eliminating AI. “And they need to issue a statement, not just about AI, but about an apology to those affected,” Kaufer said. “This ‘apology’ from Threads seems disingenuous, saying the app is ‘playable’ as if it somehow excuses racist/sexist/intolerant quotes.” In response to this incident, Kaufer decided to delete his Fable account.
So did Trammell. “The right course of action would be to disable the feature and carry out rigorous internal testing, including newly deployed safeguards to ensure, to the best of their ability, that no other platform users are exposed to harm,” he said.