TechTalk: UK ‘lagging behind’ with R&D tax relief in AI ‘Wild West’

The insurtech and startup broker world is a hotbed for innovation.

But as with much innovation, firms with great ideas can often fall short of funding – which can hamper productive businesses getting off the ground.

One way that startups can ease the financial strain of developing new products, processes and services is to use research and development (R&D) tax credits.

When utilised in the insurance sector, the purpose of R&D tax credits is to provide tax relief to an investment made by an insurtech, start up or early-stage broker that has spent money on R&D.

The government scheme is intended to ensure that firms are not put out of pocket by allowing them to claim back some or all of the money spent on R&D

But the insurtech sector has indicated that the R&D tax credit system is unfit for purpose, leaving it in the lurch as it tries to take advantage of a coming artificial intelligence (AI) revolution in technology.

Back in March, the spring budget saw a partial reversal of planned cuts to (R&D) tax credits announced.

While any forward momentum for increased R&D funding could be welcomed, the lacklustre implementation plans outlined in the budget disappointed many insurtechs and startups.

Using the spring budget’s proposals, eligible insurtechs spending more than 40% of their outgoings on R&D would be able to claim back £27 from the Treasury for every £100 spent on R&D.

One such firm who has benefitted from R&D tax credits is digital-first broker and insurtech Hubb.

Its chief executive Mark Costello told Insurance Times: “If it wasn’t for R&D tax credits in the first year of trading, we wouldn’t have got to where we are.

“But, [the insurance industry] is lagging way behind our European counterparts.”

Overall, Costello waited nine months to receive R&D tax relief. This length of time spent waiting to be reimbursed can have a detrimental financial impact on a business and make the proposition unattractive for potential applicants..

Wild West AI

The announcement of reforms to the R&D tax credit system in the spring budget came alongside a fresh investment of £100m from the government into a generative Artificial Intelligence (AI) Taskforce.

This is the sort of willingness to support R&D, that the sector has been crying out for, but the process of actually applying for the funds remains difficult for insurtechs and startup brokers.

In February 2023, Aioi Nissay Dowa Insurance Europe (AND-E) and its parent firm Aioi Nissay Dowa Insurance (AND-I) made significant investment in the Oxford University-based AI firm Mind Foundry to launch a new joint venture named the Aioi R&D Lab – Oxford.

The lab aims to develop solutions to tackle significant challenges and high-stake problems in society and is dedicated to investigating how machine learning – a branch of AI – can help insurance firms better understand and underwrite risk.

Brian Mullins, chief executive at Mind Foundry, said: “The high stakes applications are where AI is used to make decisions that affect people’s lives directly, or when those decisions are made at the scale of a population.”

Michael Kainzbauer, group chief executive at AND-E, said: “It was important to find a partnership that addresses the right ethical components.”

AND-E and Mind Foundry are aiming to apply a governance framework to AI in an ethical way, “instead of just going Wild West with what AI can do”, Kainzbauer added.

However, regulatory requirements for insurance firms involved in R&D can be tricky to navigate, especially where generative AI is concerned.

A government whitepaper published in March 2023 laid out the government’s expectations around AI in financial services to the existing regulators – the FCA and PRA – by issuing five principles to ensure its ethical and safe use.

The release of this whitepaper represents the pre-regulatory stage for AI in financial services, meaning that there is little clarity for those investing in R&D to depend on.

Speaking on the potential impact of this regulation, Kainzbauer said: “Regulation is always a two-sided story”.

He continued: “Too much regulation doesn’t always help all sides as it kills innovation, but some is good.”

Kainzbauer explained that the government should instead apply regulation to protect the consumer in terms of AI.

Mullins added: “There’s a lot more opportunity for AI and insurance, but it has to be done in the right way.”

End game?

For Chris Kenning, chief executive at Stubben Edge, the insurtech players operating in the market now are much more aware of the regulatory landscape.

He explained: “Picking the winners is really hard because no one really knows what the technology can do. From my perspective we are jumping to the end game before we’ve been through the initial stages of how we actually commercialise it.

“With something complex and new, the government’s agenda to regulate it is going to require quite a fast moving and technical understanding. I think that will be difficult, particularly in new and emerging markets, because there’s not a mature skill set.”

For Kenning, the ethical policing of AI would be most tricky.

He added: “It’s a little bit like cyber risk, it’s more about education and awareness. You still need to have those checks and balances that you would otherwise have.”

A spokesperson for Insurtech UK said: “As with other sectors however, [AI] must be regulated effectively to ensure that it is used ethically and fairly, and that the benefits are shared equitably among all stakeholders. It’s imperative that AI regulation remains flexible in order to generate growth within the insurance sector but vigorous enough to instil confidence that it won’t lead to mistrust from a consumer perspective and fear from an industry standpoint. 

“Additionally, if AI algorithms are not properly designed or regulated, they could result in bias or discriminatory decision-making that negatively impacts certain individuals or groups.”

A mismatched pace

Kenning has been through the process of applying for R&D tax credits himself.

He explained: “The compliance burden of submitting R&D and the uncertainty makes it much less attractive.”

“The technology will move quicker than the regulation. But we’ve sort of jumped straight to the regulation. So, from my perspective, there’s much more [opportunity] for us to foster some kind of entrepreneurial environment in the UK.

“Our policy should be about training, education and awareness, rather than about regulation, which I don’t think is going to be the stymieing block, because that regulation will move slower.

“[AI] will effectively become unregulated for a while as the technology moves ahead of the regulatory landscape. That’s just a natural process when new technologies emerge and that’s why I think the US is better [than the UK].”

Kenning said he believed that generative AI would have “huge implications” for the insurance sector, the wider UK and the world economy.

But if the UK insurance sector is to take advantage of the massive potential of AI, further R&D support from the government would be most welcome.