"AI" plus "your client's home address and photos" is a fair thing to be nervous about. Inspectors are right to ask where the data goes, who can read it, and whether it is being fed into some model's training set. This page answers those questions plainly, and points to the actual legal documents that bind the answers — because a trust claim you cannot verify is just marketing.
Does MoldMind use your inspection data to train AI models?
No. Your inspection data, photos, voice memos, and reports are not used to train AI models. That is stated directly in the AI Disclosure, and it extends to the AI provider: the model provider (Anthropic) processes your content to generate the report but does not use that customer data for model training (see the AI Disclosure and Subprocessors pages). The data is processed to produce your report, not harvested to improve someone else's product.
There is one nuance worth stating honestly: MoldMind does learn from your corrections to improve your own reports over time, per your account. That per-account learning is about matching your voice and standards on future jobs — it is not the same as contributing your data to a shared model's training set.
What does the AI actually see, and what does it do with it?
It sees the field data you upload, and it uses it to draft reports — nothing more. When you upload photos, voice memos, lab PDFs, and notes, the AI sorts the photos, extracts the structured data, and drafts the assessment, the remediation protocol, and the client letter. What it does not do is make final determinations or send anything on its own: the AI does not decide your findings for you, and reports are never auto-sent (see the AI Disclosure). The inspector reviews and approves every report before delivery — AI-assisted, not AI-generated.
This boundary is deliberate and it is the honest one. The AI handles the assembly; you own the judgment and the sign-off (EPA, Mold Remediation in Schools and Commercial Buildings frames the inspector-owned assessment).
Who processes your data besides MoldMind?
A short list of named sub-processors, each disclosed publicly. The Subprocessors page names the third parties that handle data on MoldMind's behalf — the AI provider (Anthropic), and the infrastructure, authentication, payment, and email services that run the product. Naming them is the point: a vendor that will not tell you who touches your data is asking you to trust a black box.
If a sub-processor changes, the subprocessors list is where it is recorded. That transparency is what lets you actually evaluate the privacy posture instead of taking a slogan on faith.
How do you verify any of this for yourself?
Read the documents, not the marketing. The claims on this page are bound by the Privacy Policy, the AI Disclosure, and the Subprocessors list — those are the authoritative texts, and this article is a plain-language summary of them. If the article and the policy ever disagree, the policy governs. That is the right relationship between a friendly explainer and a binding document.
The reason to care, beyond your own comfort, is your client's. You are handling photos of someone's home and, sometimes, health-related context, and your E&O exposure includes how you handle that data (see documentation for E&O insurance). A report tool whose data handling is documented and verifiable is part of doing that responsibly. See the three reports every job needs and the sample report.
Try MoldMind free — 3 jobs, no card.
Sources
- MoldMind, Privacy Policy — the binding data-handling terms.
- MoldMind, AI Disclosure — what the AI does and does not do; inspection data not used to train models.
- MoldMind, Subprocessors — the named third parties that process data.
- EPA, Mold Remediation in Schools and Commercial Buildings — the inspector-owned assessment the AI assists with.