Which country will win the global AI race? And will Covid19 create new winners and losers?

We’ve explored four post-pandemic futures — and what they mean for the global AI race

BY ALEXANDRA MOUSAVIZADEH, ALEX CLARK, LUKE GBEDEMAH AND ROBERTA SGARIGLIA

This article is part of Tortoise’s pre-read content for the upcoming Global AI Summit on Friday 15 May. It’s open to all — register your place here.

Just a few days to go how until the #GlobalAISummit, a fantastic online event which Tortoise is hosting. Throughout the day an array of speakers including Thomas L. Friedman, NY Times columnist, Pulitzer Prize winning author and Helle Thorning-Schmidt, former Prime Minister of Denmark to name but two will examine the key issues surrounding the impact that Covid19 has had, and will have on the development of Artificial Intelligence.

For me perhaps the most interesting sessions could be the one at 10AM to discuss AI as a national strategy: will the race continue as before? Among those taking part are Nigel Toon, CEO Graphcore, John Sawers, Former Head of MI6 and Sana Khareghani, Head of UK Government Office for AI

Before the outbreak the AI race was focused on which countries had the lead on things such as quantum computing, facial recognition and deep learning. Who had the strongest talent pool to innovate and deploy AI and who had the best strategies that ensured government funding and the most conducive operating environments.

So how has Covid19 impacted this race? Will this alter national priorities away from A.I. preparedness towards building capacity to be prepared for black swan events? Shift front-end innovation to the health care sector? Will China and the US remain locked in the quest for owning the edge on AI capabilities or will smaller nations emerge in the new world order.

This article looks in great depth at these issues presenting a quarter of possible scenarios.

Don’t forget to join us at the Global AI Summit on Friday May 15th from 9AM. Registration is free

This article is part of Tortoise’s pre-read content for the upcoming Global AI Summit on Friday 15 May. It’s open to all — register your place here.

If the 1960s had started with a global pandemic, would an American still have been first on the Moon?

Perhaps the crippling economic shock of lockdown would have forced the White House to choke off Nasa funding. Or maybe, witnessing New York’s hospitals at breaking point, too many would-be astronauts would have opted for careers in medicine. Maybe the American public, desperate for a way out of self-isolation, would have had little patience for policymakers shooting for the stars.

Right now, there’s no space race for Covid-19 to turn upside down. But, as the first wave of the pandemic hit, a new geopolitical technological tustle was underway among the major powers of the 21st Century. It is one that now faces a highly uncertain level of disarray: the global AI race.

Where were we headed before the pandemic?

A dominant United States, an encroaching China, and clusters of nations jostling well behind — that was the clear picture that emerged from our Global AI Index, a ranking of 54 countries on their development and deployment of artificial intelligence technologies that we launched in December 2019. In the race to advance capabilities and claim a share of the global market, the US seemed to be unassailable in its lead. Although rapidly progressing and buoyed by a huge population, widespread data aggregation and an increasing dominance in AI academia, China was still a long way behind.

Here’s a snapshot of the table, but for the fuller picture go here.

Within just a few months, this new world order our Index uncovered could be falling apart. A global recession eclipsing that of the 2008 financial crisis may snuff out the booming AI industry that helped boost the US to the top of our ranking. A need to collect personal data on an unprecedented scale is causing a crisis of faith for a privacy-focused Brussels, threatening Europe’s shot at a world-first policy framework for AI. A widespread need to enforce lockdown measures for an extended period may take away China’s lead on surveillance tech.

But the pandemic is also providing opportunities for AI nations. Healthcare systems are overwhelmed, making the prospect of automated disease detection and drug discovery all the more enticing. Our locked-down lives are accelerating digitisation in both business and education, providing a genuine chance for an AI-enabled virtual workplace and classroom. The open source community has flooded the internet with coronavirus-related data and code — there have never been so many data scientists in so many places all working on the same thing.

Artificial intelligence was already viewed as an area of vital national interest before the crisis. More than 42 countries had published “national AI strategies” before December 2019. We combed through them all, compiling the number of AI projects, promised funding and timeframes. Chances are, in light of the coronavirus crisis’s upheaval, all of these strategies will now need a rewrite. So what post-pandemic scenarios will AI policymakers have to contend with in the coming months and years? And where will these take the AI race?

Our four futures

On top of the toll that the virus is taking on public health, and the staggering economic impact of the lockdowns, the international response to Covid-19 is also creating enormous uncertainty. For businesses around the world, the measures intended to mitigate the spread of the virus are also causing widespread decline, and the plans for recovery are not yet clear. The threat of a second or even third or fourth wave of infections is paralysing governments.

In a recent dialogue with Salesforce, Deloitte analysts put forth a range of scenarios describing the “world remade” by Covid-19. We have considered the implications of these possible outcomes for the global AI landscape — from renewed battles over semiconductors to increased demand for surveillance technologies — and how The Global AI Index might look different in the years to come.

Future I: Business as Usual

How the pandemic unfolds:

• Despite a slow start, the pandemic is contained, and a deep recession soon gives way to economic recovery
• This effective response is driven by international collaboration, both between governments and within the scientific community
• Businesses still feel the bite of a recession, hitting the smallest the hardest and driving the biggest towards cost cutting

How this impacts the AI race:

• The global market for artificial intelligence contracts initially, except for healthcare, but recovers strongly, backed by businesses looking to cut costs through automation
• Pools of talented practitioners remain in work, avoiding the widespread job losses of other sectors
• Against the backdrop of a global scientific response to Covid-19, the trend towards international collaboration on AI research accelerates.
• The US maintains its lead, its economy stalling but recovering
• US-China relations improve, and both are secure in their positions, although China continues to challenge American supremacy

The seismic impact of the economic lockdown and the pandemic itself is going to cause lasting change. But an internationally coordinated, science-led response involving health systems around the world working alongside politicians may lead to a slow but steady return to normal.

In this scenario, the long-term fiscal impacts of the pandemic are severe and governments are forced into debt, but the recession is short-lived and recovery sets in during the latter part of 2021 as confidence returns. The short-term contraction in investment and prospecting on innovation causes a slow-down in the AI sector — something we’re already beginning to see in our data.

Meanwhile, the world’s scientific community collaborates on coronavirus at an unprecedented level. The pandemic changes the way the world does science — and AI is no exception. Our data pre-Covid showed that AI papers jointly published by authors from different countries were rising — under this scenario, we might expect such collaboration to not only continue but accelerate. At first, this is driven by proactive and preventative measures in public health, as AI researchers take advantage of new relationships forged in the crisis, but, as recovery sets in, transnational research groups expand further into other areas.

Once the brunt of the crisis is borne, businesses quickly look to make efficiency gains, and those holdouts that have not yet moved towards automation and AI solutions begin to do so. This drives rising demand for AI systems and practitioners capable of building them. Small-to-medium size businesses do not cope nearly as well with the prolonged lockdown and many face insolvency. Smaller, highly innovative AI companies face similar adverse conditions. Larger companies, on the other hand, cushioned by their capital reserves, are able to attract talented AI practitioners.

Kept afloat by the activities of its massive technology firms, the US maintains its position as the dominant AI powerhouse throughout the pandemic. Alongside this, China’s industry is accelerated and President Xi JinPing continues to pursue development as a means of solidifying Chinese military and economic security.

Future II: A New Digital Age

How the pandemic unfolds:

• The pandemic is severe and causes recession, but governments falter in their response and must turn to the private sector, concentrating power in the hands of a few businesses large enough to survive
• Private-public partnerships drive the recovery, with the private sector sharing health expertise and data
• Tech giants in particular step up to the plate, bringing a new wave of digitisation in business and education as societies continue to live in lockdown

How this impacts the AI race:

• The AI sector flourishes as people all over the world embrace digital technology more extensively — driving demand for AI products from virtual assistants to biometrics
• The US maintains its lead through its large tech giants, some of which will hoover up technological immature competitors, becoming more powerful
• Trend towards international collaboration on AI research continues, but is coordinated more than ever by companies
• The US pulls further ahead in terms of capacity, as some of its largest companies further tighten their grip on AI technology

Despite monumental efforts by national governments to coordinate responses to the pandemic, and mitigate the economic damage, they are overwhelmed. Large, technologically mature companies step in to become part of the solution. An uptick in public-private partnerships signals the dawn of a new age of company based leadership on health issues, community building and public safety. Our societies become more accustomed to tele-working and online networking than ever before.

In this scenario, the movement of large tech companies to the fore places AI at the heart of not just the recovery, but also the society remade by Covid-19. The sector experiences short-term growth in areas such as computer vision, virtual reality and other systems that can enhance lives still isolating from the outside world. Online security is also a growth area, driven by demand for new and improved communications technology, and people become more comfortable sharing data with companies.

Without an effective intergovernmental response, lockdown measures remain in place longer than in our first future. As a result online education becomes a major component of the new norm. We have already found massive online open courses are seeing a huge uptick in traffic prompted by the high numbers of people isolated at home — we would expect these trends to continue.

As the private sector takes the lead, corporate leadership and research institutions collaborate more extensively. The number of contributions from international researcher teams rises as a share of the whole, and major institutions collaborate with companies from overseas. US research into artificial intelligence is already more corporate than its competitors. We have found that corporate affiliation is nearly three times as common in research papers published in the US, indicating that the American research environment would not face much disruption in transitioning to a new digital age.

Throughout this scenario, immature companies do not survive the lockdown and resulting downturn. This causes a spree of hiring amongst large companies of the remaining AI practitioners. The global skills shortage does not improve, but the talent is concentrated even further onto the staff of massive corporations. Online learning, taking place at unprecedented levels after the pandemic subsides, helps maintain the world’s total number of data scientists.

The US is already home to a major reservoir of specialised talent. A recent study by Element AI found the share of global AI experts in the US to be around 50 per cent, meaning that growth in the sector will disproportionately benefit the country. In addition, whilst the lion’s share of AI specialists are currently part of smaller tech companies, the new digital age might see this balance tip towards the giants more likely to survive the pandemic unscathed. Here, the US stands to benefit massively from the activities of its largest technology companies. The likes of Facebook, Microsoft and Alphabet become an even more crucial component of the US economy, boosting their capacity to develop and deploy artificial intelligence.

Future III: An Eastern Sunrise

How the pandemic unfolds:

• The pandemic is still severe, but recovery from it is not uniform all over the world
• There is a big disparity between the success of measures to deal with the pandemic in the East and West
• China fills the vacuum on the world stage left by the US and Europe reeling from prolonged lockdown and economic collapse

How this impacts the AI race

• The rankings shift dramatically as China rises ahead of trajectory and the US begins to decline
• A new geopolitical bloc forms within which China becomes the engine of AI development, orbited by smaller nations like Hong Kong and Singapore
• Digitisation still accelerates, but is driven more by Chinese tech giants than American counterparts
• The US lead falters as rising unemployment, fiscal catastrophe and a culture of isolation due to the severe pandemic drives AI talent away
• US and China relations worsen, with 5G and semiconductors becoming a key battleground in a renewed trade war

In terms of managing the Covid-19 pandemic, a marked difference between the east and western hemispheres becomes clear. Western nations struggle to contain the virus, and to facilitate recovery from the prolonged economic shutdown. This contrasts sharply with the East, where consistent, highly regimented and efficient controls lead to less economic and social decline. Led by China, South Korea and Taiwan, the global epicenter of political power moves eastwards.

In this scenario the use of artificial intelligence to contain and combat the pandemic is widely regarded as a fundamental good. Moreover, the industry and companies able to take part in the provision of public health solutions, testing technologies and forms of population tracking take centre stage. In China, as well as other Asian countries, the AI sector booms, and the export of Asia-made systems reaches an all time high.

The Covid-19 outbreak has put unprecedented pressure on policymakers. Many grand plans have fallen by the wayside — including Washington and Beijing’s “phase one deal”. Struck in January after 18 months of talks, the trade deal aimed at pausing the two superpowers’ descent into trade warfare through agreements on intellectual property and technological transfer.

But by late March, the chairman of Huawei was warning that the Chinese government would retaliate to US measures curbing access to semiconductors made using US-designed tools. “Phase one” is now on ice, with tensions exacerbated by the global pandemic and the diverging paths the two nations have taken in tackling the virus. As the Chinese economy emerges from shutdown, and the situation in the United States deteriorates further, we’d expect the relationship between the two nations to worsen under this scenario.

China is three years into its national AI strategy, the New Generation Artificial Intelligence Development Plan — an initiative intended to transition China into a leading role in the global artificial intelligence industry. Through government funding, private investment, research projects and a supportive policy framework, the plan has already made China a powerhouse of AI development. China’s rise is being piloted by some enormous companies that are challenging their Western counterparts. Tencent’s WeChat, Huawei and Xiaomi, Alibaba and JD.COM are all entering the echelons of the world’s fastest growing and most valuable businesses.

Throughout our Index data, we found clear evidence of China closing the gap with the US. Under this scenario, we’d expect China to further extend its lead in AI research and begin to match American commercial funding for AI.

In 2019, Joy Dantong Ma, an analyst at the Paulson Institute, warned that “if American loses its openness edge, then the country risks pushing AI talents right back in the arms of its competitors, including China”. In this scenario, that edge might be well and truly lost. Practitioners may not be as attracted to the US market, and China may be able to mitigate some of its disadvantages in terms of top talent. In addition, technical standards and internationally accepted guidance on ethics and regulation may be increasingly influenced by Asia.

Future IV: Island Nations

How the pandemic unfolds:

• Governments everywhere fail to cope with the pandemic, causing an economic catastrophe that pushes nations towards isolationism
• Global growth stagnates while trade tensions sharpen, particularly between China and the US, setting the stage for production chain segmentation and decreased collaboration, rendered worse by the current crisis
• Both East and West suffer during waves of pandemics, and no resoundingly successful strategy for coping emerges

How this impacts the AI race:

• Surveillance and tracking technologies are prioritised as governments hold out for a vaccine, while data protection regulations are bypassed to allow mass data sharing
• AI sector becomes key to developing national security solutions
• Overall capacity for developing AI falls off dramatically as researchers withdraw from international institutions and funding dries up
• The rankings of the Index are largely unchanged, but raw scores, and intensity fall
• Reduced mobility across and within nations as a result of security measures and lower generalised trust leads to less innovation, information-gathering and collaboration, three fundamental prerequisites for AI

In the wake of a prolonged pandemic, and serial waves of outbreaks around the world, the economic and societal impacts of the crisis are dramatic. National restrictions on activity of almost all kinds create a sense of isolation within communities and amongst previously global businesses. Countries become increasingly resistant to foreign interaction, as well as infringing further and further upon personal freedoms in order to tackle the virus. Island nations close their borders, tightening their internal lockdown controls, and global supply chains break down alongside diplomatic relations.

We have already started observing signs of this: the WTO reported that, as of 22 April, 80 customs territories and countries had set restrictions on exports in place. Under this scenario we’d expect trade to descend into tit-for-tat, with retaliations moving to more strategic sectors, including AI. Further still, as the free movement of goods grinds to a half, the global recession bites even harder. With Covid-19 infections peaking at different times in different locations, trade’s pivotal role in moving supplies from where they are abundant to where they are lacking will be lost.

The development of artificial intelligence in most countries outside of the US, China and UK is dependent upon outside collaboration and trade of inputs. Even in these leading nations, capacity would fall, as key supply chains and information sharing networks unravel. From the trade in semiconductors, to the sharing of data, disaggregation would be harmful to every nation’s ability to develop AI. In addition, the dangerous arms race dynamics already playing out in the technology battleground between the US and China would likely intensify. AI companies focused on military and security technologies would likely experience increased demand, but their individual ability to innovate will fall at the same time.

US and China imports, exports and deficit all declined as a result of the Covid-19 epidemic according to the US Bureau of Economic Analysis, which entails less overall supply chain integration across countries, higher costs of production for nations that heavily rely on “global specialisation”, and therefore generalised economic contraction.

This could mean that importing and exporting essential hardware for the deployment of AI such as semiconductor chips may be disrupted, significantly limiting countries’ AI capacity development. The Semiconductor Industry Association reported recently that global semiconductor sales decreased by 2.4 per cent in January 2020 relative to the same month the previous year, and disaggregated data showing changes in the 3 month moving average for the different regions seem to show a consistent picture.

In addition, AI is a human capital- and information-intensive industry, which cannot thrive without the free flow of data and people. As a result of lockdown measures, as well as reduced economic activity, general movement of people to workplaces and places of transit has dropped significantly across all countries and continents already — such movement levels would be the new normal.

Increasingly isolated nations would not promote international flows of talent, and hiring would slow. Alongside this online education might accelerate but the top-level exchange of knowledge necessary for leading-edge AI would be stifled. International conferences, hackathons and conventions would be a thing of the past.

Where now?

The future is shrouded in an unfamiliar type of uncertainty. Whilst some elements of these scenarios may seem more or less likely than others, it is important to be aware of them all. The race to claim a technological edge in artificial intelligence is likely to be a major determinant of the geopolitical world order in the years to come.

Back in 2017, Russian President Vladimir Putin pronounced that “whoever becomes the leader in artificial intelligence, becomes the ruler of the world” — a statement that should resound on the world stage. In 2020, and in light of the Covid-19 pandemic, its significance is undiminished. The open-ended competition of our fourth future, with its severe arms race dynamics, threatens more than just our economies, but perhaps even our species.

This is one of the reasons why The Global AI Index, and the measurement of changing capacity for artificial intelligence on a global scale is so important. Tortoise Intelligence is working to develop more comprehensive metrics for progress in the areas of ethics, regulation, collaboration and diversity in the artificial intelligence sector. We are exploring these questions, and many others, at the Tortoise Global AI Summit on Friday 15 May 2020.

This article is part of Tortoise’s pre-read content for the upcoming Global AI Summit on Friday 15 May. It’s open to all — register your place here.