Tamil Nadu calls for probe into alleged public purchase of 'AI' books: an unproven political accusation

By UrbanAcres (with reporting from The Times of India) · November 8, 2026.

The School Education Department of Tamil Nadu, in southern India, has filed a complaint with the Directorate of Vigilance and Anti-Corruption (DVAC) asking it to investigate the alleged purchase of fraudulent, AI-generated books. The books were reportedly bought for the Periyar Arivulagam library in Coimbatore and for other public libraries under construction in the state. School Education Minister A. Rajmohan announced the complaint and called the case the state's "first AI scam."

The minister said the review of the purchase began after writer Perumal Murugan, a member of a government committee, raised concerns about the books. Rajmohan says the irregularities go beyond a single library. He also announced a special committee to examine public library collections and said the government will seek to recover the money if irregularities are proven.

These are the details the minister laid out. All of them are his own statements, not the findings of an investigation:

• Emails to the publisher, Agal Vilakku Pathippagam, were sent on February 14 with a seven-day deadline to deliver the books. The purchase order was issued in March. • The books were chosen from a list submitted by publishers and reviewed by a 137-member expert committee. According to the minister, 112 members attended the meeting and signed the minutes. • He wants scrutiny of the publisher's use of different addresses for purchase and delivery, and of the fact that it runs two or three publishing companies. • He says the current TVK government did not order the purchase and only released pending payments so the library could open. He blames the previous DMK government, which, according to available reports, has responded by also calling for a DVAC investigation.

The case carries a strong element of political rivalry. Rajmohan is demanding an apology from the DMK and suggests that ministers in the previous government may have influenced the approval, without ruling out officials. He also claims that author names such as "Vijay Trisha" were deliberately invented to mock a political leader. For now, these allegations come with no public evidence. He says the June date stamps on the books show when they were unpacked and placed on the shelves.

It is worth stressing what we don't know. The available material offers no independent analysis showing the books were written by AI, and no amounts of money or specific titles. "AI-generated" is, for now, the government's label. The only verifiable facts are the complaint to the DVAC and the announcement of the committee.

Our take: even if the case turns out to be an episode of partisan fighting, it points to a real, near-term risk. Generative AI makes it cheap to produce books that look legitimate: covers, authors and text in bulk. That lets a dishonest supplier pad a catalog at almost no cost. Public procurement, which relies on publisher lists and committees reviewing dozens or hundreds of titles, is an obvious target. More broadly, we are already seeing the same pattern in fake reviews, filler publications and document fraud.

The answer is not to ban AI. It is traceability: checking that the author exists, that the publisher is who it claims to be and that someone has actually read what is being bought. It is the same audit logic we demand of AI agents in business: rapid adoption without controls creates gaps that someone ends up exploiting.

In the long run, the same tools can help detect this kind of fraud by cross-referencing publisher records, addresses and text patterns at a scale no 137-person committee can match. The transition requires institutions that adapt. We will have to wait for the DVAC and the special committee to separate provable fraud from political ammunition.