E-book platform admits using AI translation without disclosure
An e-book platform offering classical literature using AI translation failed to disclose this to users despite earlier claims of manual editorial review. After users discovered the undisclosed AI use, the operator apologized, offered full refunds, and pledged future transparency.
What happened
E-book subscription service μ± λλ (Book Thief) began service at the end of 2025 offering over 700 classical literature titles in digital form for 19,800 won, claiming users would own them permanently.
Users questioned the rapid growth of titles and low translation quality, suspecting AI use. The operator initially claimed staff spent nights reading and editing works directly and received translation review from a published novelist.
On 25 July 2026, the operator admitted via social media that AI translation was being used, stating "we don't just throw AI translation together" but refined it through collective effort. Users discovered this disclosure had never appeared in the payment interface, terms of service, or platform information.
The operator had previously stated: 'AI translation is low level so we spend nights directly reading and editing works repeatedly' and 'we receive translation review from a novelist who won a literary award.'
Reported impact
Users expressed breach of trust, with one stating 'if I had known it was AI-translated, I would not have purchased.' The operator raised subscription prices from 19,800 won to 49,000 won starting 1 August 2026.
The operator offered full refunds to users who paid before 27 July 2026 and promised to transparently disclose AI use and translation process across all service channels and content going forward.
Index analysis
The platform deployed AI translation to render classical literature from public domain sources without prior disclosure to consumers, contrary to earlier explicit claims of manual editorial work.
Governance controls
After user backlash, the operator issued a public apology and announced full refunds for affected users and a commitment to future transparency about AI use.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.