Tax due diligence on an acquisition has traditionally focused on well-understood risks: historical tax positions, open audits, transfer pricing arrangements, and the profile of the target's intercompany dealings.

But as artificial intelligence (AI) becomes increasingly embedded in how businesses operate, generate value, and make decisions, acquirers and their advisers need to ask a new set of questions - ones the existing tax frameworks were not designed to answer.

In our first article, AI and tax: Implications for global businesses in Australia, we explored how tax practitioners are using AI in their work. In this article, we consider what the introduction of AI means for the application of tax rules in due diligence.

The core problem: AI and the mismatch with traditional tax concepts

As previously discussed, the Australian tax system, and the international system it sits within, are inherently human-centric frameworks that have largely focused on physical presence. But as AI systems can now operate, learn, and generate value with minimal or no human intervention, it creates a mismatch with established rules on permanent establishment, transfer pricing, and tax residency.

For an acquirer conducting due diligence, this mismatch creates risk that may be overlooked. A target that has AI at its core — managing customer relationships, setting pricing, optimising supply chains, or developing new algorithms — may have inadvertently created tax exposures that neither it nor its advisers have fully considered.

As we have explored in AI and tax: Implications for global businesses in Australia, current tax frameworks struggle to address this issue. The tax frameworks were not designed to consider autonomous systems in businesses, and it remains to be understood how AI adoption will or should affect the taxation of profits. Furthermore, existing control frameworks used to manage and identify tax risk may not be appropriate where AI has been integrated into business systems.

Key Due Diligence Risks

1. Transfer Pricing and Intangibles Migration

AI systems – including algorithms, models, training data, and the systems themselves – clearly constitute valuable intangible assets under transfer pricing principles. A target that has developed, enhanced, or migrated AI-related intangibles across borders without adequate transfer pricing documentation may be sitting on significant dormant risk.

The ATO has been proactive in this space with the release of PCG 2024/1.

However, as we have previously explored, it is yet to be fully understood whether the documentation required by the ATO under that framework fits with AI usage. Arguably, it is inherently designed around human activity and therefore each organisation would need to specifically assess their operations to determine how they could document their AI integrated business model.

During due diligence, an acquirer should critically assess whether the target has adequate documentation of its AI-related DEMPE functions. But what does this look like in practice? Should the absence of such documentation be treated as a red flag, regardless of whether the target has historically received ATO scrutiny?

2. Corporate residency and central management and control

Under Australian law, a foreign-incorporated company can be treated as an Australian tax resident if it conducts business in Australia and has its central management and control here.

As boards and executive management increasingly look to AI to improve their decision-making, it is important to consider whether some high-level policy and strategic decisions have effectively shifted to AI.

Where a target's AI systems are hosted on servers in a different jurisdiction from where its directors and management are located, an acquirer must consider whether those systems are effectively exercising central management and control, and if so, what that means for the target's tax residency profile.

3. Reliance on existing tax controls and governance frameworks

Most corporations now have robust tax governance frameworks which incorporate tax controls. However, these frameworks have been designed on the basis of systems that are inherently controlled by humans and have not yet been adapted to consider, document or control risks associated with AI. If AI decision-making is now being incorporated into aspects of the accounting and finance systems, like accounts payable, what does this mean for the target's tax risk profile? It might mean that their existing tax governance framework is deficient and adds little assurance, despite its best efforts.

What should acquirers do?

Due diligence enquiries in this area should go beyond standard checklists. At a minimum, an acquirer should seek to:

• Map the target's AI ecosystem — understanding what AI systems the target uses, where they are hosted, what decisions they make autonomously, and what value they generate.
• Review intercompany arrangements — assessing whether the target's policies properly reflect AI-related functions and whether intercompany agreements are aligned with actual operations.
• Assess documentation adequacy — given the increasing autonomy of AI, it is critical to determine where "control" actually sits, particularly where AI is now making decisions previously made by senior executives.
• Consider residency risk — particularly for targets with cross-border structures that rely on AI systems hosted outside the jurisdiction of the nominal decision-makers.
• Factor in regulatory uncertainty — international tax rules are likely to develop more specific guidance as AI deployment accelerates, and it is reasonable to expect that rule changes will follow.
• Review controls – identifying what frameworks govern automated decision-making and how these manifest in the target's financial and accounting systems.

AI represents a fundamental challenge to international tax frameworks built around human decision-making and physical presence. A target's AI footprint is no longer just a technology or operational consideration; it is a tax risk that demands careful analysis.

The cost of failing to ask these questions during due diligence could significantly affect the economics and the risk profile of the deal. Contact our team to discuss this in more detail.