Loan application forms and AI data leakage: what a home loan application gives away in ChatGPT or Claude.

The loan application form is where a broker turns a household into a data record. It holds two identities, employers, income, every account and card, the children, the property and the deposit. Brokers paste it into ChatGPT or Claude to extract it into a lender portal, to check it for gaps and to turn a rambling fact find into the tidy version a credit assessor reads. This page marks every item on a typical home loan application form that the National Consumer Credit Protection Act and the Privacy Act bear on, and shows how the same form is processed on Ermos inside the brokerage.

The loan application form

Completed with the applicants after the first meeting and before anything is lodged. Section 117 of the National Consumer Credit Protection Act requires a licensee to make reasonable inquiries about the consumer’s requirements, objectives and financial situation, and to take reasonable steps to verify that situation, before making the preliminary assessment.

Home loan application: client details and declaration
Harbour Ridge Finance Pty Ltd · Confidential · signed 15 October 2026

Applicants
Contact and residence
Employment and income
Dependants and expenses
Assets
Liabilities
Loan requested
Deposit
Declaration
Broker

An invented application form. Licence, passport, phone, email and account values are placeholders, and the credit union is invented. Hover or tap any highlighted item for the analysis; the full list is below.

What leaves the brokerage

  1. 1
    Applicants and identity documents

    Two complete identities with the government identifiers used to verify them. Personal information, and the material identity theft is built from. Whether the Privacy Act reaches a particular brokerage depends on its turnover and activities (section 6D); the clients expect confidence either way. Privacy Act guide.

  2. 2
    Address, rent and contact details

    Where the family lives, what they pay, and the phone and email a scammer impersonating the broker or the lender would use. Privacy Act guide.

  3. 3
    Employers and income

    Two named employers, start dates, base pay and overtime. Financial information the brokerage collects because section 117 requires reasonable inquiries about the consumer’s financial situation and reasonable steps to verify it. Privacy Act guide.

  4. 4
    Children and declared living expenses

    Two children, one in a childcare arrangement, and the clients’ own estimate of what they spend. Part of the financial situation the broker must inquire into, and a figure later tested against bank statements. Privacy Act guide.

  5. 5
    Assets and account endings

    Savings, super and a vehicle, with the endings of two accounts. A balance sheet and enough identifiers to make a convincing call to the bank. Privacy Act guide.

  6. 6
    Liabilities

    Card limits and balances, a car loan with a named credit union and a student debt. The commitments a lender counts, and a map of the household’s credit relationships. Privacy Act guide.

  7. 7
    The loan and the property

    The address, price, loan amount and LVR. The deal itself, which a competing buyer, an unscrupulous agent or a fraudster posing as the conveyancer would want. Privacy Act guide.

  8. 8
    A gift from a relative

    The source of the deposit and the donor’s name. A third person’s financial affairs, collected so the gift can be documented for the lender. Privacy Act guide.

  9. 9
    Declaration and consents

    What the applicants agreed to: collection, disclosure to lenders and credit reports. Whether that consent reaches an AI provider is the question the brokerage has to answer before pasting the form. Privacy Act guide.

  10. 10
    Broker and licence

    Public register details. Low exposure on their own, and they tie the document to the licensee that owes the best interests duty under section 158LA.

What happens in Frontier Models (ChatGPT, Claude)

  • The whole document is transmitted to the provider’s servers. On personal plans the user cannot choose where it is stored or processed, and for the major providers that has generally meant outside Australia, under that country’s legal process. Some business plans offer storage in Australia, but the provider still receives the content.
  • On consumer tiers, prompts can be retained and used to train future models.
  • The provider’s staff and systems can reach the content under the provider’s own policies.
  • Once sent, the brokerage no longer controls it. Retention and deletion happen on the provider’s terms.

The same document on Ermos

Ermos Dominion

The document is drafted on a unit in your office, from your own templates and the client file. Every marked item stays on the unit, which keeps working with the internet switched off. Nothing in it is transmitted to an AI provider, so there is no disclosure to analyse: the confidentiality work is the same as for your file server.

Ermos Edge

Your documents, the index and your access boundaries stay on your own instance in an Australian data centre. When a staff member asks for this draft, only their question and the excerpts they are permitted to see go to a named Australian provider, who is contractually barred from keeping the content or training any model on it. Nothing is processed outside Australia.

Frontier Models (ChatGPT, Claude)Ermos EdgeErmos Dominion
Where the document is processedProvider servers, generally offshore on personal plansYour instance in an Australian data centre; answers generated in AustraliaThe unit in your office
Who can read itProvider staff and systems, under policyYour staff, within the boundaries you set. The named Australian provider processes only permitted excerpts, under contract, and keeps nothingYour staff only
Is it keptOn the provider’s termsKept on your instance; the provider is barred from retaining itStays on the unit
Can it train a modelOn consumer tiers, yesNo, by contractOnly your own unit learns from it, on your hardware

Part of AI data leakage in mortgage brokerages. See how this works in practice: Application packaging on Ermos.

Common questions

Do ChatGPT and Claude use the information in a home loan application form?
Anything pasted from the form goes to the AI company's servers and is processed there to produce a reply. As at October 2026, whether a conversation on a personal plan (ChatGPT Free, Plus or Pro; Claude Free, Pro or Max) can be used to train the models depends on a setting each user controls. Business plans (ChatGPT Business and Enterprise, Claude Team and Enterprise) and the developer APIs sit under commercial terms, and OpenAI and Anthropic each say they do not train on that content by default (OpenAI, Anthropic). On either product, the thumbs up and thumbs down buttons can send a whole conversation back for training. One form carries two identities, a driver licence and a passport number, two children and a $60,000 gift from Sophie's father.
Is it safe to paste a home loan application form into ChatGPT or Claude?
No, because the form holds two complete identities and a household's whole financial position. The Office of the Australian Information Commissioner (OAIC) recommends in its guidance on commercially available AI products that organisations do not enter personal information, and particularly sensitive information, into publicly available generative AI tools, because of the significant and complex privacy risks. Taking out the names still leaves the employers, the Kingswood address, the Glenmore Park purchase price and the account endings, and together they describe one family. Removing names rarely makes a document anonymous. Information stays personal information while the person is reasonably identifiable, and the OAIC warns that information can be at risk of re-identification even when it has been de-identified or anonymised.
Can I upload a home loan application form to ChatGPT or Claude instead of pasting it?
An uploaded form leaves the brokerage as a whole file, which can carry comments, tracked changes and file properties as well as the part you meant to share. The OAIC's guidance also tells organisations to review carefully the terms and settings that will apply to their use of any AI product. For this form, the identity numbers, the declaration signed on 15 October and the details of the gift from Gerald Pryor would all travel in the same file.
How can a mortgage brokerage use AI for a home loan application form without sending it to ChatGPT or Claude?
Ermos is built so the document stays under your control. On the Dominion AI unit it is drafted inside your office and none of it is sent to an AI provider. On Ermos Edge your documents and index stay on your own instance in an Australian data centre, and only the question and the excerpts the user is permitted to see go to a named Australian provider that is contractually barred from keeping them or training any model on them. The draft form is produced from the client file and the brokerage's own application template. The broker checks every entry against the clients' documents and signs it off.
Does the Privacy Act apply to every mortgage brokerage?
No. A business with annual turnover of $3 million or less is generally a small business operator and outside the Act, and commission income counts toward turnover. Exceptions in section 6D(4) include disclosing personal information about someone for a benefit or service without their consent. Many brokerages are covered. Even those outside the Act owe clients confidence and answer to their licensee and aggregator.
What does the law require a broker to ask on the form?
Section 117 of the National Consumer Credit Protection Act requires reasonable inquiries about the consumer's requirements and objectives and about their financial situation, and reasonable steps to verify that situation. It does not prescribe the form. The form is the broker's record of the inquiries, which ASIC's RG 273 expects brokers to keep.
How long does the brokerage need to keep the form?
There is no single period. Section 120 lets a consumer ask for a copy of the preliminary assessment for up to seven years from the date of the quote, and ASIC's RG 273 says retention depends on factors such as the loan term. APP 11.2 requires an APP entity to take reasonable steps to destroy or de-identify personal information it no longer needs, unless the law requires it to be kept.

See Dominion answering from your own documents.

A demonstration takes 30 minutes, uses no client data, and comes with no obligation.

Book a demo