Modi government will use AI to detect fraudulent investor claims
What's the story
The Indian government is gearing up to use artificial intelligence (AI) technology to detect fraudulent claims made by investors for unclaimed shares, dividends, and matured deposits. The National e-Governance Division (NeGD), under the Ministry of Electronics and Information Technology (MeitY), plans to deploy 12 AI services for this purpose. The move will help the Investor Education and Protection Fund Authority (IEPFA) verify applicants' identities before processing refunds.
Fraud detection
NeGD has identified 28 known and potential fraud patterns
The NeGD has identified 28 known and potential fraud patterns in a request for quotation (RFQ) document.
These include cases of "forged and altered identity documents" and entitlement letters being re-uploaded after alterations.
The authority has also flagged cases of multiple claims filed against the same securities, same mobile number linked to multiple PANs, and same email address appearing across unrelated claims.
Advanced detection
Combining document checks with analysis across claims
The RFQ also notes newer forms of suspicious behavior, such as "LLM-generated templated 'reason for claim' text at volume."
This refers to repetitive claim explanations that could have been generated using large language models.
The proposed system will combine document checks with analysis across claims and verify information against official records wherever possible.
It will also match names across records and identify claims linked by shared identifiers.
Operational details
AI system will suggest relevant noting codes
The AI system will prepare a synopsis of each claim and suggest relevant noting codes used by the authority.
However, it is important to note that the final decision will be made by a government officer.
The RFQ clearly states "the score is informative," with action on a claim resting with the back-office officer.
Project timeline
Implementation period set for 4 months
The NeGD has set a four-month implementation period for the AI system, followed by an initial 12-month operations and maintenance period.
This can be extended by two more 12-month periods if needed.
The RFQ was published on October 13, with proposals due by November 5.
This project marks a significant step toward using advanced technology to combat fraud in investor refund claims.