Will ARIA Sing?

The much trailed UK version of ARPA now has a name, and it’s not BARPA or UKARPA, it’s ARIA: the Advanced Research and Invention Agency. Not, note, Innovation but Invention. Is this going to be an important distinction or simply permit the old trope of ‘Brits are good at inventing but not making money’ to come to the fore again? Before this week’s formal Government announcement, the Commons Select Committee on Science and Technology had already published its Report on the proposal, initially announced in the 2019 Queen’s Speech. If I were to paraphrase their report, I would say this group was a bit suspicious of what the agency might be for and what it might look like, stating:

“It is strange that more than a year after its inclusion in two successive Queen’s Speeches, the Government has not clearly articulated the need for, or intended remit of, the proposed agency. To date, it seems to be a brand in search of a product.”

And, indeed, the announcement this week still leaves wide open what the remit will be. With £50M initially on the table (out of a committed total of £800M in the years ahead), it is clear that it will need specifically to identify one or two target areas, at least in the first instance; the field cannot be left undefined and it would be unwise for the money to be salami-sliced across different challenges. However, to date there is no ‘customer’ explicitly identified, unlike DARPA in the USA for which the Defence Department (and hence the D in its name) is key, so what areas are likely to feature is totally unclear. The remit is still to be narrowed down.

The intention is that this agency will be distinguished from other funding sources by focussing on ‘high-risk, high-reward scientific research’. In other words, just like DARPA as spelled out carefully in the book by Sharon Weinberger The Imagineers of War, failure will be a necessary part of ARPA’s success. If there are no failures, then the funding decisions will turn out to have been too conservative. That attitude does not necessarily apply more widely!

However, aside from not having specified the remit of ARIA, there are also, as yet, no details of how it will operate structurally. DARPA has operated by giving programme leaders – selected for their vision and technical expertise – much freedom in selecting the teams that get the funding, by bringing together researchers (academic and industrial) into groupings to fulfil some identified and ambitious goal, a moonshot as it were. So, the two key ingredients are visionary programme managers and funding delivered to teams brought together for a specific purpose. It is not about funding individuals (as with much UKRI funding, or the ERC), nor funding long-standing collaborations. Innovation may require significant disruption to break some persistent logjam or to think the unthinkable.

This world of ARIA should recognize the importance not just of failure – and be willing to embrace it – but also of this disruption. And at this point it seems to me it is key to recognize that this means diversity in the broadest sense. Boardrooms around the world are waking up to the fact that diversity can lead to better profits because moving away from group-think, the traditional behaviours of an established customer base and managers who may all resemble each other, it is possible to bring new perspectives that open up products and markets. A US study last year of diversity and inclusion among S&P 500 companies concluded

“Diverse and inclusive cultures are providing companies with a competitive edge over their peers.”

How does this translate into innovation in science and technology research?

The evidence here is that diversity – in the sense of encompassing the work of minorities ­– again plays a major role in innovation.  A study in PNAS of over a million US doctoral theses (so here it is the diversity background of individuals which is being considered, not teams) over four decades showed that underrepresented groups produce higher rates of scientific novelty, according to the specific methodology used. This, you might think, would lead to highly successful future careers for these minorities, but this was not the outcome, as is probably rather obvious. On the contrary, novel outputs from racial minorities and women (typically the minority gender) were less likely to be cited and built upon by others. It has been well documented that women’s research gets fewer citations (see e.g. last year’s report from the Royal Society of Chemistry on gender bias in publishing) and the penetration of the work – for instance to lead to innovation in products – will therefore also be reduced, with subsequent consequences for careers. The PNAS study suggests this is equally likely to be the case for racialised minorities, with ensuing negative consequences for their careers too. However, not only their careers, but innovative developments, exactly what ARIA intends to aim at.

So, the evidence suggests, if you rely on the dominant group – typically white men – to drive innovation, you may be making a fundamental mistake. I hope that as ARIA gets going, it will create structures which enable those who make decisions about who gets the money to be visionary in many different ways, not least in looking beyond what one could call the ‘usual suspects’. This isn’t simply a matter of justice; the evidence suggests innovation will be more likely to arise if a wealth of perspectives is brought to bear on whatever target(s) are selected for funding rounds.

It will be interesting to see how the ARIA landscape unfolds, as the plans for this potentially exciting new agency crystallise and as – it is to be hoped – it finds, to quote the Select Committee, the product of which it is the brand.