Artificial intelligence (AI) is helping to drive innovation in every sector. Indeed, the scope and potential for AI-powered research and development is huge, not only helping to accelerate innovation in virtually every sector you can think of, from drug discovery to predictive logistics, but also by fuelling the development of the foundational AI-powered tools that make these exciting breakthroughs possible.
In this article, we take a look at the barriers that could potentially slow AI-take-up, and explore what’s driving investment in artificial intelligence R&D in the UK. We also explain how innovators can benefit from R&D Tax Credits to help fund their work.
Challenges for artificial intelligence R&D
The UK Government’s AI Opportunities Action Plan identified several “barriers” and other challenges that could hinder AI adoption.
Some of the main challenges include:
- The UK doesn’t yet have the AI infrastructure in place to meet the needs of either the public or private sector. More must be done so that organisations across the whole of the UK can access a sufficient supply of computational power (either through investment at home or by creating international partnerships).
- Data is the “lifeblood” of AI, but there are several barriers that currently hinder the use of high-quality public and scientific datasets. For example, many high-value datasets are private, making access difficult due to intellectual property rules. In some areas, the government is sitting on so much data it’s not even sure of how useful or valuable it is. In others, using the data for AI training could raise privacy and ethical concerns. There are also worries that data isn’t being collected strategically, or that when it is collected it’s not of sufficient quality for use.
- There’s currently a technical AI skills gap in the UK workforce (“tens of thousands” more AI professionals are needed by 2030), which, if not solved, could seriously slow down the UK’s ability to develop and adopt AI.
- Many people are uncomfortable with their data being used for AI training, and steps need to be taken to foster public trust, addressing very valid privacy and security concerns. At the same time, AI regulation needs to be balanced with a legislative environment that is pro-innovation, as poorly designed regulation could otherwise unwittingly throttle progress.
The UK Government appears to recognise the seriousness of these challenges, which is why they’ve written a dedicated “AI Action Plan” to pave the way for businesses to develop and implement creative and powerful new AI solutions (see below).
Drivers for artificial intelligence R&D
Perhaps the biggest driver of AI adoption is that most businesses recognise the vast potential of the technology, and some are already enjoying the benefits. A recent study from SAP and Oxford Economics showed that UK businesses plan to increase their investment in AI by 40% over the next two years. The study reported that businesses are already seeing a 17% return on their AI spend, and ROI is expected to nearly double to 32% within the next two years.
The opportunity to boost productivity is so impressive that the UK’s Modern Industrial Strategy has identified artificial intelligence (AI) as one of the core “frontier technologies” with both high growth potential and an expectation that it will play a key role in “profoundly reshaping our economy”.
The strategy lays out the UK Government’s plan to boost public investment in R&D, citing our “research and engineering talent, vibrant start-up and scale-up scene, frontier companies, and global leadership on safety and governance” as the ideal foundation for fostering innovations that embrace the opportunities of AI.
Following on from their strategy, they’ve created an AI Action Plan that aims to add £47 billion annually to the economy by boosting productivity in key sectors such as healthcare and financial services.
To support this, the UK Government has set up a £500 million Sovereign AI Unit to fund UK startups and help them grow. They’re also boosting local infrastructure with last year’s launch of the Isambard-AI supercomputer in Bristol and the designation of five AI Growth Zones designed to fast-track data centre construction.
Plans are also progressing for the launch of a new National Data Library to “speed up medical research, improve the efficiency of public services, and help British firms develop new AI‑driven products and services.”
R&D Tax Credits are available for artificial intelligence R&D
AI projects within innovative companies are unsurprisingly on the rise as more businesses adopt the technology to stay competitive. Any business considering undertaking such a project should consider applying for relief, as their work may very well qualify as R&D for tax purposes.
This is because AI integration projects will invariably need a significant investment of effort to successfully bridge the gap between new AI and legacy systems. Most projects will need to undertake deep data analysis, bespoke development, and rigorous systematic testing to ensure the new technology actually works as hoped.
For the integration to work, AI automation requires extensive engineering to overcome compatibility challenges, ensuring that data can flow between old and new architectures without any issues. Some projects will seek to deliver physical
automation, introducing robots that engineers must design, prototype, and test. This is complex work that requires genuine scientific or technical advancement, which is precisely why these activities often qualify for R&D Tax Credits.
For projects that qualify, companies may be able to claim relief on a range of R&D-related costs, including staff costs, software, data licences, cloud computing, contractor payments, consumables and prototypes.
Examples of eligible artificial intelligence R&D activities include:
- Developing new AI algorithms to detect diseases earlier from scans.
- Creating AI systems that can predict when roads need maintenance (e.g., spotting potholes).
- Building novel Large Language Model (LLM) applications, like chatbots, where the integration requires solving unique technical challenges (e.g., accuracy, security, or data bias).
- Developing “sovereign” AI capabilities (i.e., building new, foundational models rather than just using off-the-shelf APIs).
- Engineering ways to merge legacy systems with new AI automation.
For a project to be eligible for R&D Tax Credits, it must involve more than just using AI. It needs to develop or significantly modify the technology to overcome a scientific or technical challenge (including complex mathematical challenges) that a professional in the field could not easily solve.
Our clients’ success stories
- Precursive: Precursive is a provider of AI-powered Professional Services Automation (PSA). By partnering with Leyton, they successfully claimed R&D Tax Credits and finished their year ahead of budget. The benefit they received helped them to bring new hires forward earlier than originally planned.
- Tanridge Capital: Tanridge Capital uses data-driven, algorithmic models to make trading decisions systematically, taking the emotion out of decision-making based on market data and risk management. Their work is complex, so they partnered with Leyton to help them educate HMRC on how their highly technical R&D projects qualified for tax relief.
If you enjoyed this article, you might also like:
