AI-Generated Receipt and Invoice Detection

AI-generated receipt and invoice images are an emerging fraud pattern in promotions and expense workflows. When Tamper Detection is enabled, Taggun analyses supported submissions for signals associated with AI image generation. This capability is part of Digital Tampering Detection, not a separate endpoint or response family.

Availability

📘

Availability

AI-generation analysis is part of Digital Tampering Detection, included on the Advanced plan and above. On verbose extraction endpoints, enable Tamper Detection in your account Feature Settings when it is available on your plan. On Receipt Validation endpoints, also set fraudDetection.allowTamperDetection: true for the campaign. See Plans & Feature Access.

How It Works

1. Submit — send a supported document through an eligible verbose extraction or Receipt Validation endpoint.
2. Analyse — Taggun runs the enabled tamper analysis, including checks for signals associated with AI image generation.
3. Respond — the result is returned through the existing tamper-detection response.

Internal analysis methods may change and are not part of the public API contract. Build against the documented response fields rather than a particular detection technique.

Request

There is no AI-specific endpoint or request parameter. On verbose extraction endpoints, the check is attempted when Tamper Detection is enabled for the account; a populated result is not guaranteed for every document. For Receipt Validation, configure the existing fraudDetection.allowTamperDetection campaign setting; see Campaign Settings Reference.

Response

AI-generation analysis does not create a separate public response shape. Read the result through the existing tamper fields documented in Digital Tampering Detection.

When present, isTampered is Taggun's recommendation and tamperedScore is the associated numeric signal. Optional detail fields are informational and may change. A returned tamper result does not by itself prove that a document was AI-generated or establish the submitter's intent.

Best Practices

📘

Digital does not mean AI-generated

A legitimate emailed receipt or invoice may be digital without being AI-generated. Digital Receipt Detection classifies whether a submission appears digitally created or photographed; AI-generation analysis belongs to Tamper Detection.

• Route flagged submissions according to your risk policy, with human review for consequential decisions.
• Combine tamper analysis with relevant checks such as Duplicate & Similarity Detection rather than treating one signal as proof of fraud.
• PDFs and standalone images are processed differently. Do not assume every image-level check applies identically to PDF content.
• If your team confirms an AI-generated or fabricated document that passed detection, follow Contacting Support.

Need something not covered here? Contact [email protected] — see Contacting Support.