Skip to content
HRAreceipt.in

Anti-Fraud · 31 May 2026

Spotting AI-Generated Rent Receipts: 7 Tells

AI-generated rent receipts fail seven repeatable tests under Rule 26C scrutiny. Combined with a single Annual Information Statement reconciliation, an HR team rejects an AI bundle in under two minutes.

In this section

Answers

View all 97 articles →

What are the seven tells of an AI-generated rent receipt?

Short answer

AI-generated rent receipts fail seven repeatable tests under Rule 26C[1] verification: identical landlord signatures, calendar-impossible dates, PAN-stamp ink mismatch, copy-paste address blocks, round rent amounts, malformed PINs, and ITR boilerplate phrasing. A single Annual Information Statement[8] reconciliation closes the rest.

Through 2026, AI image generation has caught up to template fidelity on any single-receipt scan. What it has not caught up to is the cross-receipt and cross-system footprint a real twelve-month tenancy leaves on the AIS under Section 285BA[3]. Tells 1-7 are the fast filter; AIS is the conclusive filter; the verification QR short-circuits both.

  • Identical landlord signature across twelve months. Real landlords sign with natural variance: slope, pressure, spacing. AI bundles match pixel-for-pixel across every receipt. Two side-by-side resolves it.
  • Calendar-impossible or single-day-cluster dates. Receipt dated 31-Apr, 29-Feb in a non-leap year, or all twelve receipts on the same day-of-month. Real landlords issue when they collect; 2-4 days drift month-on-month is normal.
  • PAN stamp in different ink from the body. Real PANs are stamped or hand-written in the same ink as the signature line. AI bundles render the PAN in a different font weight or colour the model invents as a stamp aesthetic.
  • Copy-paste address blocks across allegedly different landlords. Multi-year submissions naming three landlords show identical PIN-format, identical commas, identical floor-prefix conventions. Real landlords format addresses idiosyncratically.
  • Rent amounts ending in identical round thousands across all twelve months. Real rent carries deposit-adjustment paise or rent-revision arithmetic. AI bundles produce twelve receipts at exactly Rs 25,000.00 each. Cross-check against the rent agreement and the bank UTR trail.
  • Missing or malformed PINs. Indian PINs are exactly six digits. AI bundles drop them, produce five-digit codes, or generate PINs that do not resolve on India Post pincode search[7].
  • ITR boilerplate inside the receipt body. Real receipts say *received from*. AI receipts inject *for HRA exemption under Section 10(13A)* or *for Form 12BB submission*. Landlords do not think in tax-code vocabulary; only the tax filer does.

How does AI output differ from an HRAReceipt-generated receipt?

Short answer

Across all seven tells, an AI bundle and an HRAReceipt-generated receipt produce structurally different artefacts. The table maps each tell against AI behaviour, HRAReceipt generator behaviour, and the verifier cross-check that resolves it.

TellAI bundleHRAReceipt generatorVerifier cross-check
Landlord signaturePixel-identical across 12 receiptsSingle landlord-name field; signature line stays for pen-and-paper anchorTwo receipts side-by-side
Date stampsCalendar-impossible or single-day clusterDate picker rejects 31-Apr, 29-Feb non-leap, future datesForm 12BB month declaration
PAN blockDifferent font or ink from receipt bodyPAN-format validator (AAAAA9999A) on submit; structured fieldIT Dept PAN Verification API + AIS reconciliation
Address blocksCopy-paste style across allegedly different landlordsTyped once per landlord; PIN validated separatelyIndia Post pincode + cross-receipt comparison
Rent amountIdentical round thousands across all 12 monthsPer-month amount field; honours paise; deposit-adjustment line supportedAIS bank-channel rent line + Section 269ST cash cap
PIN codeMissing, 5-digit, or non-resolving6-digit validator on submit; rejects malformed PINsIndia Post pincode lookup[7]
Receipt body phrasingInjects tax-code vocabulary (Section 10(13A), Form 12BB)Static "received from {tenant}" pattern; tax-code phrasing absentRead the body line aloud

Each tell mapped against AI behaviour, HRAReceipt-generated behaviour, and the verifier cross-check that resolves it.

What single cross-check beats every visual scan?

Short answer

The Annual Information Statement under Section 285BA[3]. Bank-channel rent above Rs 50,000 per month appears against both tenant and landlord PAN. A bundle claiming Rs 60,000 per month while AIS shows nothing, or the reverse, is rejected without scanning a single signature.

From FY 2025-26 the AIS rent line ingests bank-channel rent above Rs 50,000 per month via NEFT, IMPS, UPI, and RTGS. The mismatch engine runs automatically inside the Faceless Assessment system under Section 144B[6]. AI fabrications cannot create AIS entries. Cash-rent bundles separately face the Section 269ST cash transaction limit: aggregate receipts above Rs 2 lakh trigger a 100 percent penalty on the recipient. AIS and 269ST together leave AI bundles nowhere to hide.

What short-circuits the entire scan?

Short answer

A receipt carrying a verification QR. The QR binds to document content at the moment of legitimate generation; any subsequent edit invalidates it. A two-second scan confirms authenticity and skips tells 1-7 for that document.

AI fabricates pixels. AI cannot fabricate a valid verification QR. For HR teams processing 200 Form 12BB[5] submissions per quarter, the QR path takes 10-20 percent of incoming receipts off the manual scan entirely. The seven signs of AI-generated salary slips covers the salary side; the typed-vs-handwritten audit under Rule 26C covers the broader receipt-validity question.

Generate a Rule 26C-clean rent receipt with QR

What is the 3-minute rent-receipt verification workflow?

Short answer

Six steps. QR scan first, fifteen seconds. If absent or invalid, run the visual scan against the seven tells, then pull the Annual Information Statement[8] for any flagged file. Total: under three minutes per receipt at sustained volume.

The workflow front-loads the cheapest checks. QR scan short-circuits 10-20 percent of bundles entirely. Visual scan catches another 60-70 percent at the surface. AIS reconciliation under Section 285BA[3] is the final filter for residual files. At 1,000-file daily HR volume, the cadence keeps throughput intact.

Decision-tree flowchart of the 3-minute HR verification workflow: an incoming Form 12BB submission first hits a QR scan; if the verification QR is present and validates, the receipt is accepted and HRA exemption is granted under Section 10(13A); if absent or invalid, the verifier runs a visual scan against the seven tells (signature, dates, PAN, address, amount, PIN, phrasing); if any tell is flagged, the verifier pulls the Annual Information Statement for bank-channel rent reconciliation; AIS mismatch under Section 285BA routes the file to rejection with Section 144B Faceless Assessment risk; AIS reconciliation grants HRA exemption.
The decision tree. Source: Rule 26C IT Rules 1962 · Section 285BA AIS rent line · Section 144B Faceless Assessment.

References

  1. 1.Rule 26C, Income Tax Rules 1962 — Employer evidence requirement when annual rent exceeds Rs 1 lakh.
  2. 2.Section 10(13A), Income Tax Act 1961 — House Rent Allowance exemption framework.
  3. 3.Section 285BA, Income Tax Act 1961 — AIS reporting basis. Rent line expansion from FY 2025-26.
  4. 4.Section 269ST, Income Tax Act 1961 — Cash receipt aggregate cap of Rs 2 lakh; 100 percent penalty.
  5. 5.Form 12BB (Rule 26C statement) PDF — Form fields include PAN or Aadhaar of landlord above Rs 1 lakh.
  6. 6.Section 144B, Income Tax Act 1961 — Faceless Assessment scheme; auto-routes low-structure inputs to manual review.
  7. 7.India Post Pincode Search — Official PIN code lookup; six-digit validation.
  8. 8.Annual Information Statement (AIS), Income Tax e-Filing portal — Taxpayer AIS access; bank-channel rent line reconciliation.

References & related

Primary sources

Related guides