View from Washington: Aukus looms over AI and quantum
The new US-UK-Australia alliance is set to shake up how all three countries carry out research in key emerging technologies.
Most of the talk has been about submarines, but another important aspect of the new Aukus alliance between Australia, the UK and the US is that it defines emerging technologies – particularly artificial intelligence and quantum computing – as first-order national security issues.
As Tom Tugendhat, chair of the Commons’ Foreign Affairs Committee, said in a Twitter thread:
“Bringing together the military industrial complex of these three allies together is a step change in the relationship. We’ve always been interoperable, but this aims at much more. From artificial intelligence to advanced technology the US, UK and Australia will now be able to cost save by increasing platform sharing and innovation costs. Particularly for the smaller two, that’s game-changing.”
Tugendhat is right. The game has changed, and in ways that are only just coming to light. For example, digital innovation has been driven by communications, e-commerce, consumer electronics and the PC since the mid-1980s, even though the sector originally depended on the defence industry. This alliance puts government and security back at the forefront.
In other ways though, Aukus reflects a consolidation of how the technology landscape has evolved during the last five years amid greater competition between China and the West and recurrent talk of ‘decoupling’.
As well as restricting the US activities of several Chinese companies through its Entities List (most notably, but not exclusively Huawei), Washington has blocked Chinese-led takeovers of companies it considers particularly sensitive, such as Lattice Semiconductor. Aukus itself is then consistent with the recommendation of March's US National Security Commission on AI, chaired by former Google chief Eric Schmidt, that the US needed to not just increase its own efforts but also "rally our closest allies and partners to defend and compete in the coming era of AI-accelerated competition and conflict".
UK regulators are still mulling over US company Nvidia’s proposed $54bn bid for Cambridge-based Arm partly for national security issues – more on that later – and Prime Minister Boris Johnson has launched probes into a clutch of others. These include a Chinese-backed deal for semiconductor manufacturer Newport Wafer Fab and one with suggested Chinese involvement involving the takeover of a Welsh graphene specialist, Perpetuus Group.
For its part, China has hardly sought to hide that it views AI and other emerging technologies as key to defence as well as future economic prosperity. Its ‘New Generation Artificial Intelligence Development Plan’ was published in 2017 and calls for the country to be the world leader in the sector by 2030. It became a policy priority following the 2016 defeat of the world Go champion, Lee Sedol, in a tournament against AlphaGo, an AI developed by UK-based Google subsidiary DeepMind Technologies. China’s military leaders see Go as an important proving ground in the development of strategic thinking for the battlefield.
The landscape has changed greatly since, also in 2016, Theresa May’s government nodded through Softbank’s original acquisition of Arm, back then passing over concerns raised in the Ministry of Defence similar to those being taken more seriously today.
With the AI race well under way – whether you like it or not – the consolidation within Aukus of the research efforts of the three countries promises not only the technological benefits Tugendhat identifies but also feels like a necessary acceleration.
But there will be a price.
The cycle for delivering consumer and other branches of civilian innovation has shortened from the 18 months in Moore’s Law to one that is now, to all intents and purposes, annual. Of late, defence applications have often used the benefits of programmable logic to which various secret combinations of spices and sauces would be added. However, it has been clear for a while that the worlds of hardware and software are eliding for AI, and quantum computing will require a shift to entirely new architectures. As a result, what companies can and cannot release to the public, and when, is likely to come under much tighter official scrutiny. Time-to-market vs. Defence of the Realm(s).
Consolidation as well as greater cooperation across the three countries is also a possibility, and this brings things back to Arm-Nvidia. As two world-class companies operating in the technology spaces covered by Aukus, and given the environment the alliance seeks to create, it may be much harder for UK regulators to block the deal. Indeed, they may now want to encourage it. Meanwhile the EU, which has serious antitrust concerns over the union of the leading IP provider with a leading chipmaker, may feel – understandable French anger notwithstanding – that there is too much political risk in objecting, particularly with some members nervous about the extent of President Joe Biden’s commitment to Nato.
Then, some of the more notable consequences may be for the global research infrastructure, one that had become increasingly freewheeling since the fall of the Berlin Wall.
Some familiar voices are already proclaiming Aukus as evidence of the ‘Brexit dividend’. Never mind the facts that technological collaboration between the three members is already taking place through the Five Eyes intelligence alliance (with New Zealand and Canada, both not part of this agreement); that the US and UK have been sharing the nuclear propulsion research covered since 1958 and already overlap hugely in defence research (e.g., BAE Systems); and that the technological and national security trends in AI and quantum have only surfaced since the referendum vote (‘Let’s spend £250m a week on R&D’, anyone? Anyone?)
That said, as emerging technologies are considered more sensitive, governments are going to reconsider how far they can go in undertaking certain types of cutting-edge work through multinational economic bodies like the EU and other civil partnerships rather than military alliances operating under strict secrecy. Just how ‘open’ exchanges in technical conferences covering those areas can be in future is also now even more up for debate.
These issues have always been there. And they have always been tricky. But are we at a point where they are about to be as tricky as they were half a century ago, and when those who knew how to navigate such territory have either retired or passed away? And, of course, we do not yet know where any boundaries are going to be set.
Many in the UK technology sector will see Aukus as a great opportunity. They are probably right to do so. But, even if not entirely in public view, the three powers involved need to communicate how they expect commercial and academic collaboration to work clearly and, given the alliance positions the areas within its scope as pressing and serious, quickly. Submarines may look like item one on the agenda, but everything else is equally immediate.
Once you’ve shook things up, you must still reorder them. Ad hoc simply isn’t an option. The months ahead will be busy ones. Well, they better had be.
Sign up to the E&T News e-mail to get great stories like this delivered to your inbox every day.