AI Usage Disclosure

No AI processing in v1

ErgoTax v1 does not use artificial intelligence or machine learning models as part of normal product operation. Specifically:

• Tax documents you upload are not processed by AI models
• Field extraction uses deterministic PDF text parsing, not AI or OCR
• Tax calculations are performed by a deterministic Rust engine, not an AI model
• No document content or tax return data is sent to AI service providers

What the processing pipeline actually does

When you upload a document, ErgoTax uses a WebAssembly build of the pdfium PDF library to extract the text layer from the file. That text is then matched against known field positions for that form type using a deterministic layout-profile system. No neural network, language model, or AI inference step is involved.

Tax calculations are performed by the ergotax-core Rust engine compiled to WebAssembly. All calculations follow published IRS and state agency rules. There is no AI involved in any calculation step.

Why this matters

Tax document processing with AI raises legitimate concerns: AI models can hallucinate, their outputs can be unpredictable, and routing tax documents through external AI services creates privacy risks. ErgoTax's position is that tax calculation and field extraction should be deterministic and local — not dependent on AI model behavior or external service availability.

Future versions

If AI features are introduced in a future version of ErgoTax, this disclosure will be updated to describe what AI is used for, what data is involved, and how users can understand or opt out of AI-assisted features.

Contact

Questions about AI usage may be sent to [email protected].