AI and the Future of Transfer Pricing: Risks and Opportunities

ARTICLE | August 18, 2025

Authored by Aprio, LLP

This article was originally published by Aprio on August 14, 2025.

At a glance

• The main takeaway: Global tax authorities are poised to increase their use of artificial intelligence to improve transfer pricing audits.
• Impact on your business: Multinational companies should prepare for increased scrutiny in some areas, carefully harness proven AI tools for increased efficiencies in their own operations, and seek the help of an experienced transfer pricing team.
• Next steps: Working with a knowledgeable transfer pricing advisor can help you respond to any audit inquiries and tackle planning, compliance, and reporting burdens with greater ease.

The full story on AI and the future of transfer pricing:

The explosive growth of artificial intelligence (AI) in the past five years has resulted in a fast-changing landscape across all industries and sectors, including tax and transfer pricing. While multinational companies may look to generative AI for help reducing their global tax rate and avoiding penalties, taxing authorities are also poised to increase their reliance on AI to flag high-risk tax returns for further examination.

AI tools for worldwide taxing authorities

Agencies such as the IRS United States), HMRC (United Kingdom), the Australian Taxation Office, the Austrian Tax Authority, and the French Tax Authority have begun integrating AI into their transfer pricing audit processes. The IRS’s 2024 IRA Strategic Operating Plan, Annual Update Supplement, for example, states that the agency is using AI to help determine audit targets, focusing on compliance risks in the areas of “partnership tax, general income tax and accounting, and international tax.”

With access to more data and experimentation with increasingly sophisticated algorithms, tax administrations can now detect patterns, anomalies, and pricing inconsistencies that might have previously gone unnoticed. Some authorities are even developing proprietary AI models to predict outcomes in dispute resolution processes like Mutual Agreement Procedures (MAPs) and Advance Pricing Agreements (APAs), potentially streamlining negotiations and improving case management.

However, the use of AI tools comes with some concern, such as regulatory compliance. Transfer pricing rules vary widely across jurisdictions, and AI tools may not be fully attuned to these nuances. A model trained on the Organization for Economic Cooperation and Development (OECD) principles might not account for country-specific regulations, potentially leading to findings of non-compliance. Moreover, even when AI aligns with global standards, it must still adhere to the arm’s length principle, a requirement that can be difficult to validate if the underlying logic of the AI is unclear.

Ethical considerations further complicate the picture. AI systems can inadvertently perpetuate biases present in historical data, such as favoring certain industries or regions. This not only affects the fairness of the analysis but could also raise reputational risks if stakeholders perceive the outcomes as discriminatory or unjust. Additionally, questions about accountability surround the use of AI tools in a scope such as the enforcement of international transfer pricing. When AI systems are used to enforce transfer pricing rules without clear accountability, it becomes difficult to determine who is responsible for errors, misjudgments, or biased outcomes. A lack of transparency can erode trust between taxpayers and taxing authorities, and may undermine the legitimacy of enforcement actions if decisions cannot be adequately explained or defended.

Multinational enterprises and the use of AI

As multinational enterprises (MNEs) navigate increasingly sophisticated regulatory environments, AI is reshaping how transfer pricing analyses are conducted, documented, and defended.

Historically, transfer pricing analysis has relied on established methodologies such as the Comparable Uncontrolled Price (CUP), Transactional Net Margin Method (TNMM), and the Profit Split Method. While these approaches are grounded in sound economic principles, their execution often involves labor-intensive processes. Many MNEs are exploring AI-driven alternatives for tasks such as data collection and classification. AI systems can extract and structure data from a wide range of sources, such as public databases, financial filings, ERP systems, and even unstructured documents like emails and contracts. Natural language processing (NLP) enables these tools to identify relevant information with significant speed and accuracy, reducing the burden on tax teams.

However, while AI offers powerful tools for streamlining and enhancing transfer pricing analysis, it also introduces a set of risks that organizations must carefully consider. One of the most fundamental concerns is data quality. AI models rely heavily on the data they are trained and fed with; if that data is incomplete, outdated, or biased, the resulting analysis is equally compromised. For example, scraping financial data from public sources may inadvertently include non-comparable entities or omit critical context, leading to flawed benchmarking.

Another significant challenge, as noted above, is the lack of transparency in many AI systems. Machine learning algorithms, especially those that operate as “black boxes,” can produce results that are difficult to explain or justify. This poses a problem in the transfer pricing domain, where tax authorities expect clear documentation and defensible methodologies. If an AI tool cannot provide a transparent rationale for its selection of comparable or pricing decisions, it may undermine the credibility of the analysis during audits or disputes.

The bottom line on the future of artificial intelligence and transfer pricing

As AI and other technologies continue to evolve, multinational companies conducting intercompany transactions should be aware of the impact these changes can have on their business. While the use of AI tools may increase transfer pricing scrutiny, over-reliance on AI as a response to audits or inquiries is no substitute for human judgment and international tax experience. A healthy balance must be struck between the use of AI to improve the analysis, with strong oversight from an experienced practitioner.

Please connect with your advisor if you have any questions about this article.