Pressure on HMRC to reveal if AI Drives R&D Tax Claim Decisions

0 Comments

The Tax Tribunal has recently issued a landmark ruling against HMRC. As a result, there is now pressure on HMRC to reveal if AI Drives R&D Tax Claim Decisions.

Pressure on HMRC to reveal if AI Drives R&D Tax Claim Decisions

Overview

HMRC have invested £millions in sophisticated IT software to identify non-compliance and close the tax gap. In fact, HMRC uses a wide ranging data analytical model designed to process and analyse massive volumes of data. 

Furthermore this software has access to a number of different databases ranging from bank accounts and pensions to Google Street View(!). The subject of HMRC's analytical tools is a topic we'll be covering in more detail in a separate post.

Given HMRC's investment in IT software, it's unsurprising that HMRC might be using AI. However, this particular case has sparked fresh controversy over HMRC's usage of private data. 

Why has HMRC been forced to reveal if AI drives R&D Tax Claim Decisions?

During Covid HMRC were arguably very lapse in conducting R&D tax claims - an area they perceive open to abuse. Therefore, the rules for making R&D tax claims have been tightened. Additionally HMRC may have relied upon AI as part of the process of their increased focus on challenging historic R&D tax claims.  

In this case the appellant had suspicions that HMRC was using AI to make  R&D tax claim assessments. They noticed unusual details in official correspondence. For example, American spellings, mismatched facts, and oddly formatted letters. They therefore submitted a Freedom of Information request to HMRC.

HMRC refused to reveal whether they had been using AI to conduct R&D tax claims. They argued that confirming their AI use would assist fraudsters and harm tax collection. This argument was supported by The Information Commissioner.

The appellant expressed concerns that HMRC's use of AI in such cases could place sensitive intellectual property at risk. Furthermore, they argued it could weaken the UK’s innovation economy. Therefore, they pushed for answers to protect taxpayers and rebuild trust.

The Tribunal was persuaded that HMRC's non-disclosure undermined public trust in HMRC's handling of R&D tax claims. Additionally, their lack of transparency, could potentially dissuade legitimate claims. This would be counter productive to the government's objectives to support the innovative sector. 

They also determined that HMRC had not been sufficiently transparent to support public debate on it's usage of AI in R&D tax compliance.

Consequently, the judge presiding over the Tribunal - Judge Alexandra Marks, determined in the appellant's favour. She stated  “the balance of the public interest lies in disclosing the information requested”. As a result, HMRC has been given until September 18 2025 to comply with this request.

Summary

If HMRC uses AI for scrutinising R&D tax claims UK businesses - particularly tech start ups, face serious challenges. Founders may think twice before submitting legitimate claims to HMRC. Additionally, misjudgments by AI could trigger unfair enquiries and involve costly disputes.

The tribunal’s decision sets a strong precedent for transparency. HMRC is under mounting pressure to reveal if AI Drives R&D Tax Claim Decisions. Furthermore, companies may finally determine how technology impacts on their claim(s).  Hopefully the tribunal's decision will shape trust, compliance, and the future of R&D incentives across the UK.

For more useful information, check out our Ebooks here

And if you'd like to know how we can help you with all of this, or with anything else, feel free to give us a call on 01202 048696 or email us at [email protected].

Alternatively, please feel free to complete our Business Questionnaire here.

About the author

Richard Baldwyn

I’ll help you legally pay less tax, using insider knowledge gained from my time as a former tax inspector—insight most accountants simply don’t have. More about Richard and the TFA team

{"email":"Email address invalid","url":"Website address invalid","required":"Required field missing"}
>