Fable app AI reader summaries contain racially biased recommendations
Fable's AI-powered year-end reading summaries generated racially biased recommendations, including advice for a Black user to read more white authors. The company removed the feature and apologized after user backlash revealed similar problematic summaries related to race, gender, sexuality and disability.
What happened
Fable, a book discussion and tracking app, used OpenAI's API to generate playful year-end reading summaries for users in January 2025. A Black user received a summary advising her to 'surface for the occasional white author' after reading mostly Black authors in 2024, while another user was asked if he was 'ever in the mood for a straight, cis white man's perspective.' Multiple other users reported similar summaries containing inappropriate commentary about disability, sexual orientation and gender.
Reported impact
Users expressed outrage on social media, with many vowing to delete the app and calling the summaries racist. Some writers said the incident made them question using the platform.
Index analysis
The AI system produced biased and offensive recommendations without adequate human content filtering or moderation safeguards. Fable's head of product admitted that oversight and offensive language filters had failed, and the company acknowledged the summaries were AI-generated without clear disclosure to users.
Governance controls
In response, Fable removed the component of the model that 'playfully roasts' users and removed all AI-generated 2024 reading summaries and two other AI features. The company implemented changes including an opt-out option and clearer disclosures that summaries are AI-generated, and issued a public apology.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.