17 November 2026

New Indian mega library rocked by scandal over 'AI-generated books'

BBC | Xavier Selvakumar, Yazhiniyan

The Madras High Court ordered anti-corruption investigators to freeze publisher bank accounts and probe state public library procurements following allegations of widespread procurement of AI-generated books. Novelist Perumal Murugan triggered the investigation after discovering nearly 1,800 suspicious titles at the newly opened 3-billion-rupee Periyar Arivulagam library in Coimbatore.

Government selection committees failed to detect false references and misattributed authorship across hundreds of purchased volumes, including works claiming economist Jean Drèze as author. Tamil Nadu School Education Minister A Rajmohan described the case as potentially India's first AI scam. State audits quickly exposed systemic vetting failures. Officials subsequently removed 523 disputed books from the historic Connemara Public Library in Chennai while expanding inspections state-wide. Over 160 writers, academics, and publishers signed a joint demand seeking procurement transparency, collection audits, and public fund recovery. The ongoing investigation highlights severe public-sector vulnerabilities when auditing low-cost synthetic publishing.

Comment

Generative text models expose structural vulnerabilities in state procurement pipelines that rely on static metadata such as ISBN registrations rather than content authentication. Large language models like GPT-4 enable commercial suppliers to produce plausible, hallucinated literature at scale for minimal cost. Procurement review procedures designed around financial compliance lack the algorithmic filters required to evaluate machine-generated synthetic text.

This failure pattern carries direct consequences for defence intelligence and policy research repositories reliant on open-source intelligence. Without cryptographic provenance standards such as C2PA metadata, public and military archives remain vulnerable to low-cost synthetic content insertion that degrades institutional knowledge bases.

Strategic Question for Discussion
If commercial vendors deploy models like GPT-4 to generate synthetic literature at scale, how can government analytical bodies prevent unverified synthetic text from contaminating open-source intelligence repositories without imposing prohibitive manual verification costs?
The pattern suggests that relying solely on post-hoc manual review panels will prove economically unsustainable against high-volume synthetic text generation. Instead, institutional repositories are likely to adopt automated cryptographically verified ingestion pipelines using frameworks like C2PA content credentials to flag unauthenticated material before cataloguing. Until such cryptographic verification becomes standard across publishing clearinghouses, open-source analytical units will face continuous risk from hallucinated data masquerading as genuine technical or historical documentation.
Share your assessment in the comments below.
💬
Ask me
Ask Strategic Study India New chat ×