Libby introduces AI content filters as self-published book flood threatens library catalog
OverDrive's Libby app will let readers opt out of AI-generated books, audiobooks, and translations as the digital publishing industry braces for an influx of synthetic content.
Last verified:
OverDrive, the 40-year-old digital library platform behind the consumer app Libby, is preparing to let readers selectively filter out AI-generated books and audiobooks as self-published synthetic content floods the market. According to The Verge, newly appointed OverDrive CEO Marc DeBevoise announced the company will introduce opt-in controls to flag AI authorship, AI narration, machine translation, and AI-generated illustrations—giving 92,000 public libraries, schools, and universities worldwide a middle path between full adoption and complete rejection of AI-created works.
The Scale of the Problem
The urgency is real. According to The Verge, Kobo’s CEO Michael Tamblyn revealed last month that the company rejects nearly half of all self-published submissions due to AI concerns. Amazon has already begun rate-limiting self-published uploads per day since 2023 to stem the tide of low-quality synthetic books. OverDrive doesn’t accept direct author uploads but partners with Draft to Digital, an intermediary that supplies self-published titles to Apple Books, Google Play Books, and other storefronts—channels through which AI-generated books flow into libraries indirectly.
The problem’s scale underscores what Tamblyn called “a firehose” of synthetic content entering digital publishing. However, Libby’s existing catalog offers some insulation: the app’s 6 million books have been borrowed over 1 billion times, but the vast majority predate the emergence of modern large language models. DeBevoise noted that “everything before 2020 or 2022 is by definition not AI.”
The Dual-Use Strategy
Rather than ban AI outright, OverDrive is positioning itself as a transparency and choice enabler. DeBevoise told The Verge that “we need to tell people what’s available and how it was created,” while maintaining that AI offers legitimate benefits in content recommendations and localization. Libby introduced its own AI-assisted book discovery features last year—a move that initially sparked backlash—but DeBevoise framed these as separate from the human-authorship filtering.
This approach reflects an industry-wide tension: libraries serve patrons with varying preferences, and blanket prohibitions alienate readers who welcome AI-translated works or synthetic narration for accessibility. The filter lets both camps coexist within the same app.
Why This Matters
The Libby announcement signals how legacy digital platforms are adapting to the AI publishing reality without surrendering to it. For the 92,000 institutions and millions of patrons using Libby, transparent labeling preserves agency—a crucial principle as AI-generated content becomes indistinguishable from human-authored work through publication channels. For OverDrive and peer platforms, the precedent is also commercial: libraries increasingly demand visibility into content provenance as a condition of licensing and patronage. If competitors like Kobo continue rejecting AI en masse while Libby filters but permits, library procurement decisions may shift toward transparency-first vendors. The real test arrives when AI-generated books occupy 20%+ of self-published submissions—a threshold that will force Libby’s filter from a nice-to-have into a table-stakes feature.
Frequently Asked Questions
What types of AI content will Libby's filter cover?
The filter will identify AI-authored books, AI-narrated audiobooks, machine-translated works, and AI-generated artwork within the catalog.
Will filtering AI content reduce library selection?
Unlikely significantly—most of Libby's 6 million-book catalog predates modern LLMs (pre-2020/2022). However, the self-published segment is growing faster with AI-generated titles.
Is Libby phasing out AI features entirely?
No. Libby retains AI-powered book recommendations and localization features. The filter is opt-in user control, not a blanket ban on AI.