AI in Australian Tax Practices and TPB 2026 Guidance

Artificial intelligence is increasingly transforming Australian tax and accounting practices, streamlining tasks such as tax research, document processing and data analysis.

But what are the TPB requirements for using AI in tax practices?

The TPB’s TPB(GS) 55/2026 clarifies how existing professional obligations apply when registered tax and BAS agents use AI for client work. It does not introduce a separate AI compliance regime; instead, it reinforces the need for professional judgement, appropriate oversight and practitioner accountability when using AI.

For Australian tax practitioners, the message is clear: AI can support your workflow, but it does not replace professional responsibility.

AI Can Support Practitioners, But It Does Not Replace Accountability

The central principle is straightforward: AI can assist a practitioner, but it does not take over the practitioner’s responsibility.

If an AI system produces an incorrect tax interpretation, calculation, document or recommendation, responsibility remains with the practitioner. AI-generated information should therefore be reviewed, verified and considered alongside professional judgement before it is relied upon.

For tax practices, this means AI should be treated as a support tool rather than a substitute for professional oversight.

The greater the potential impact or risk of the work, the more important appropriate human review becomes.

What Does the TPB Guidance Require?

1. Review AI-Generated Information

AI systems can produce information that is incomplete, inaccurate or misleading, even when the response appears confident.

When AI is used for tax research, practitioners should verify the relevant legislation, rulings or other authoritative sources. Similarly, AI-generated client communications should be reviewed by an appropriately qualified person before being sent.

The goal is not to eliminate AI from the workflow. It is to ensure that professional judgement remains part of the process.

2. Maintain Appropriate Records

Using AI does not remove existing record-keeping responsibilities.

Where AI materially contributes to client work, practices should consider whether their records demonstrate how the work was performed, what review took place and how professional judgement was applied. Depending on the engagement, this could include documenting the information provided to the AI system, the output received, independent checks, changes made and who reviewed the final result.

There is no single AI record that every practice must maintain. Documentation should be appropriate to the nature and risk of the work.

3. Protect Client Confidentiality

Client confidentiality is another major consideration.

Under Code item 6, practitioners must not disclose information relating to a client’s affairs to a third party without permission, unless disclosure is otherwise permitted or required. Depending on how an AI product operates, the AI provider may be considered a third party.

Therefore, entering client information into an AI system should not automatically be treated as an internal activity. Practitioners need to understand who receives the information and why.

4. Obtain Client Permission Where Required

If using an AI system involves disclosing information about a client’s affairs to a third party, client permission may be required.

The guidance indicates that permission can be documented through an engagement letter, signed consent or another appropriate form of communication. Practices should therefore review whether their existing engagement documentation adequately reflects how client information may be used with AI.

A standard confidentiality clause should not automatically be assumed to address every AI-related disclosure.

5. Conduct AI Vendor Due Diligence

Practitioners also need to understand the AI products they use rather than relying only on vendor marketing claims.

Before approving an AI tool, practices should consider what information it receives, where information is stored and processed, whether data is used for model training, who can access it, retention periods, security controls, privacy arrangements, overseas data transfers and the system’s limitations.

AI vendor due diligence should be treated as an ongoing risk-management activity rather than a one-time exercise.

Your Practice May Already Be Using AI

One of the easiest things for practices to overlook is AI functionality already built into existing software.

A firm may not have formally adopted ChatGPT or another generative AI platform, yet its accounting, tax or document-management systems may already use AI for document processing, transaction classification, data extraction, anomaly detection, automated summaries, research assistance or predictive recommendations.

That makes a technology and software inventory an important starting point.

Ask your software providers whether their products use artificial intelligence or machine learning to process, analyse, generate or recommend information. If they do, the next question should be: What happens to client information when that functionality is used?

Understand Where Client Data Goes

AI governance becomes particularly important when client information is involved.

Before entering information into an AI system, practices should understand where the information is processed and stored, whether it is transferred outside Australia, how long it is retained, whether it is used to train or improve models, who can access it and whether it can be deleted.

The OAIC also recommends a cautious, risk-based approach to privacy when using commercially available AI products.

For tax practices, this reinforces the need to avoid simply copying sensitive client information into publicly available AI tools without understanding the privacy and security implications.

A Practical AI Checklist for Tax Practices

Before using AI for client-related work, practices should ask:

• Have we identified all AI tools and AI-enabled features already in use?
• What client information does each system process?
• Where does that information go?
• Have we reviewed the provider’s privacy and security arrangements?
• Is client permission required?
• Does our engagement documentation adequately address relevant AI use?
• Have we defined approved and prohibited AI use cases?
• Is human review required before AI output is relied upon?
• Are important review and verification steps documented?
• Have staff received appropriate AI-use training?
• Are AI tools included in ongoing risk and quality-management reviews?

If these questions cannot be answered confidently, the practice may already be using AI without adequate controls.

What Happens When AI Use Is Poorly Controlled?

The risk is not limited to an incorrect AI-generated answer.

Poorly controlled AI use can contribute to incorrect tax information, inappropriate disclosure of client information, privacy breaches, inadequate consent, weak documentation, insufficient professional review and poor vendor oversight.

The TPB’s position is clear: using AI does not transfer professional responsibility from the practitioner to the technology provider.

The better question for practices is therefore not simply, “Can we use AI?”

It is:

“How can we use AI within a controlled professional process?”

What This Means for Australian Tax Practices

TPB(GS) 55/2026 does not prohibit tax practitioners from using AI, nor does it establish an entirely separate set of AI-specific professional obligations.

Instead, it reinforces an existing professional principle: technology can change how tax work is performed, but it does not change who is responsible for the work.

For practices adopting or expanding AI, the practical priorities are clear:

Identify where AI is already being used. Understand how client data is handled. Review AI vendor arrangements. Establish appropriate permission processes. Train staff. And ensure AI-generated work receives suitable human review.

For accounting and tax practices looking to improve capacity while maintaining structured processes and appropriate professional oversight, NCSGX Australia provides outsourced accounting and tax support designed to support back-office operations while professional responsibility remains with the appropriate practitioner.

As AI becomes more deeply embedded in accounting and tax technology, effective governance will be just as important as the technology itself.