{"id":5334,"date":"2021-04-07T07:36:57","date_gmt":"2021-04-07T07:36:57","guid":{"rendered":"http:\/\/www.odbms.org\/blog\/?p=5334"},"modified":"2021-04-07T07:40:23","modified_gmt":"2021-04-07T07:40:23","slug":"on-the-new-tortoise-global-ai-index-interview-with-alexandra-mousavizadeh","status":"publish","type":"post","link":"https:\/\/www.odbms.org\/blog\/2021\/04\/on-the-new-tortoise-global-ai-index-interview-with-alexandra-mousavizadeh\/","title":{"rendered":"On the new Tortoise Global AI Index. Interview with Alexandra Mousavizadeh."},"content":{"rendered":"

“I think the conversation is really about a shift of where funding is coming from. Governments are spending far less than the big tech platforms. What this tells us about who owns the direction of travel of AI is fascinating. Are we now in a position where the power that the public sector was able to deploy in the past is massively outgunned by private companies and their R&D budgets?” —\u00a0Alexandra Mousavizadeh.<\/strong><\/p><\/blockquote>\n

This is my follow up interview with\u00a0Alexandra Mousavizadeh, \u00a0Partner at Tortoise Media. We talked about the new version of the\u00a0Global AI Index.<\/em><\/p>\n

RVZ<\/p>\n

Q1. In December 2020, Tortoise Media launched the Global AI Index to benchmark nations on their level of investment, innovation and implementation of artificial intelligence. Since\u00a0then, your team has been working on expanding the Index. What are the main results of this new Index?<\/strong><\/p>\n

Alexandra Mousavizadeh:<\/strong> The most striking result is China\u2019s rapid improvement. Although the gap between the top three is significant (the US is still streets ahead of China, and the UK lags far behind in third place) across our 143 indicators, China has made gains.\u00a0That rise is mostly due to a serious boost in research: yet another Chinese university joined the Times\u2019 list of top 100 computer science universities; the total number of citations from high-achieving Chinese computer science academics jumped by 67 per cent over the course of the year; the number of Chinese academic AI papers accepted by the IEEE \u2013 a body which sets AI standards and also publishes a number of influential AI journals \u2013 now out-do those by US academics by a factor of seven; and China overtook the US in terms of AI patents granted around two years ago and has been pulling further ahead ever since.<\/p>\n

China is also pulling ahead in the roll-out of supercomputers, with almost twice as many super computers as the US, demonstrating its growing threat to the US\u2019 AI supremacy.<\/p>\n

The UK has slipped in some key metrics, and its lead over its closest competitors has narrowed. Although a new AI strategy has now been announced, it\u2019s not been published yet. We have also seen British slippage across several different key parts of the framework; universities, supercomputing, research, patents and diversity of specialists.<\/p>\n

Q2. What has been added to the Global AI Index of this year?<\/strong><\/p>\n

Alexandra Mousavizadeh:\u00a0<\/strong>We\u2019ve added Armenia, Bahrain, Chile, Colombia, Greece, Slovakia, Slovenia and Vietnam: each of these countries has recently published a strategic approach to artificial intelligence on a national level, and therefore \u2018qualified\u2019 for assessment under The Global AI Index framework. That brings the number of countries assessed to 62 overall, up from 54 last year.<\/p>\n

We\u2019ve also developed all-new national AI dashboards: policy makers can monitor their national AI activity across all 143 indicators in real-time. These dashboards, the first of their kind, can also simulate the impact of policies via the target-setting feature, which calculates hypothetical ranks and scores based on chosen policy targets.<\/p>\n

Q3. What new metrics did you introduce and why?<\/strong><\/p>\n

Alexandra Mousavizadeh:\u00a0<\/strong>We\u2019ve added a range of new metrics to deepen our measurements across many of the pillars of the index. In Talent, we have incorporated data provided by Coursera, showing the level of enrollment and activity on online learning courses specific to \u2018artificial intelligence\u2019 and \u2018machine learning\u2019. This data-set fills in gaps in countries like India, China and Russia – where our other metrics were not as comprehensive – by acting as a proxy for the level of online learning taking place.<\/p>\n

We\u2019ve also refined some of our existing metrics to increase the accuracy of our data. We\u2019ve replaced the Open Data Barometer data-set, which was outdated in many respects, with measures from the OECD OURdata Index<\/a> to better reflect the level of open data use and suitability. The new source is much more recent and relevant. We\u2019ve also made our measures of 5G implementation more granular, reflecting the actual level of supported networks in a given country. This is a crucial leading indicator for the capacity to adopt artificial intelligence more widely in a given country,\u00a0 so this change offers a lot more clarity in the Infrastructure pillar of the index.<\/p>\n

In our forthcoming update for the index in May, we\u2019ll be adding more indicators on diversity issues. These will complement a component of the index that deals with regulation and ethics.<\/p>\n

Q4. Who are the biggest risers on the Index?<\/strong><\/p>\n

\u00a0Alexandra Mousavizadeh:\u00a0<\/strong><\/em>Israel<\/strong> has surged up the rankings from 12th to 5th place. It improved its talent rank, with more R package downloads, an increase in stack overflow questions and answers, and a rise in GitHub commits. These are important indicators because they speak to developments in a country\u2019s coding community beyond the formal education sector. It also still has the highest number of startups as a proportion of the population – 3 AI startups for every 100,000 people (compared with 5 startups for every million people in the US). It\u2019s an impressive feat for a small country, but its rise is driven in part by the number of proportional, or intensity-based, metrics in the index (which favour small countries), and partly by changes in our methodology that more accurately capture the number of developers and other specialists on social media.<\/p>\n

Also of note is Finland<\/strong>, which has procured a pre-exascale supercomputer and substantially increased its coding activity. Use of Python, R and commits on GitHub have grown, as well as our measures of GitHubs Stars and Stack Overflow Questions\/Answers. This is a notable result of Finland\u2019s ongoing focus on skills development, driven by both government strategy and an excellent ecosystem, including the Finnish AI Accelerator<\/a>, the Tampere AI Hub<\/a> and the AI Academy at University of Turku<\/a>. That complements exciting tech startups, including Rovio<\/a>, Supercell<\/a>, and CRF Health<\/a>.<\/p>\n

\u00a0<\/em>The Netherlands<\/strong> is the biggest riser; helped by a slow-down in some of the countries that previously ranked above them (Japan), The Netherlands have accelerated their coding and development activities and have high scores on the proportional Talent measures i.e. Number of Data Scientists per capita.<\/p>\n

Q5. How has the pandemic influenced the global development of AI in the world? Do you have any insights to share on this?<\/strong><\/p>\n

Alexandra Mousavizadeh:\u00a0<\/strong>It is difficult to untangle the exact effect of COVID on the global development of AI – a lot of the data is yet to come in. One area that has definitely suffered is start-ups. This time last year, the UK had 529 AI startups listed on Crunchbase, but it\u2019s now dropped to 338. Other countries have seen similar collapses in startup numbers. There are some counterintuitive results: the 62 countries in our Index attracted similar overall levels of private funding for AI as in 2019. Several countries have seen a fall in investment, but these don\u2019t necessarily match those that have been worst affected by the pandemic. Both the US and China, for instance, have seen a similar drop-off in funding, despite the US having a much worse pandemic overall.<\/p>\n

The pandemic has also created challenges to which some countries have responded with AI. Coronavirus has become a focal point for Israel\u2019s AI entrepreneurs, with several Israeli companies emerging as front-runners in areas like diagnostics, disease management and monitoring systems. Vocalis Health<\/a>, for example, was launched by the defence ministry and is aiming to diagnose Covid effectively based on people\u2019s speaking patterns.<\/p>\n

Q6. The lack of transparency is a limiting factor for the effectiveness of the Index. For example in the case of Russia, much of its AI spending may be going to military purposes \u2013 and you can\u2019t track it.\u00a0 What is your take on this?<\/strong><\/p>\n

Alexandra Mousavizadeh:\u00a0<\/strong>It\u2019s true that there is AI spending and research that we can\u2019t track, especially in countries that are less transparent, like Russia. We only use one proprietary data source for the Index – the Crunchbase API<\/a> –\u00a0 and the vast majority of the rest of our information is open source. Government spending on clandestine AI activity represents a small proportion of funding and progress in the AI space. Even if a country is spending quite large sums opaquely on covert AI, that spending is siloed, and often won\u2019t contribute to a country\u2019s overall progress in AI.<\/p>\n

Q7. Isn’t this also the case for most of the other counties such as China, but even the USA?\u00a0 They do not necessarily reveal their use of AI for military purposes.<\/strong><\/p>\n

\u00a0Alexandra Mousavizadeh:\u00a0<\/strong><\/em>I think the conversation is really about a shift of where funding is coming from. Governments are spending far less than the big tech platforms. What this tells us about who owns the direction of travel of AI is fascinating. Are we now in a position where the power that the public sector was able to deploy in the past is massively outgunned by private companies and their R&D budgets? Amazon spends roughly ten times as much on R&D and infrastructure as DARPA\u2019s total budget ($36bn and $3.5bn in 2019 respectively). The UK\u2019s new research agency, ARIA, is set to have just \u00a3800m over the course of the next parliament.<\/p>\n

Another related issue is that of selective publication. It\u2019s well within a company\u2019s rights to not release research that it carries out – but when that research has the potential to create massive public goods, it\u2019s concerning that decisions are made in private by unaccountable tech companies.<\/p>\n

Q9. How would you characterize the current geo-politics of artificial intelligence?<\/strong><\/p>\n

Alexandra Mousavizadeh:\u00a0<\/strong>The countries that get on top of AI will accrue enormous benefits very quickly – not just in efficiency gains and cost reductions, but in transformative technologies that will increase almost every aspect of their global competitiveness. Although a lot of research that we\u2019re currently seeing is business, not state-led, states can move rapidly if they have to.<\/p>\n

Conventional wisdom says it\u2019s a competitive environment, with some saying that we\u2019re seeing a rise in AI nationalism. But the different gains and losses made by nations in different factors and metrics on the index show a complex picture of states investing in different areas. The World Economic Forum\u2019s \u2018Framework for Developing a National Artificial Intelligence Strategy<\/a>\u2019 highlights the collaborative role that many governments are aiming to play; co-designing, rather than merely responding to, technological change across multiple sectors. And many of the factors we track don\u2019t stick to one country – talent crosses borders, and Github is global. A lot of the most interesting developments are iterated with open-source technology. So in many ways, the geo-politics of AI can be one of mutual benefit rather than a zero-sum environment.<\/p>\n

It\u2019s more complicated than any one story; but one narrative that we\u2019re definitely seeing is two superpowers, the US and China, establishing themselves as dominant in the space. Then there are a number of specialist smaller states that could have a significant role to play in terms of standard-setting, and specialist AI in areas of national comparative advantage – take the UK and medtech for example. Finally, there are countries that want to be involved but aren\u2019t currently in a place where their strategy (or level of investment) matches their ambitions. Those countries need to get serious, quickly.<\/p>\n

AI is an accelerant – we run the risk of seeing clusters of AI excellence that exacerbate divides within and between nations, compounding existing inequalities and leaving those without skills and capital behind.<\/p>\n

Q10. Carissa Veliz<\/a> mentioned that “to ensure ethical behaviour around AI, certain behaviours should be banned by law” . Do you plan to include policy regulations in the future version of the Index? <\/strong><\/p>\n

Alexandra Mousavizadeh:\u00a0<\/strong>The dashboards we developed have a set of implicit policy recommendations sitting behind the indicators; governments can see how their rankings would improve by using these features. But with regard to ethics, we spent hundreds of hours with the team discussing how regulation and ethics could feed into the index and concluded that it needed an in depth examination which warranted its own investigation. We are now doing that work. The May update to the index will contain a section on policy regulation, so stay tuned.<\/p>\n

Having said this, there are some metrics in the Index which do address ethical considerations – such as the diversity of researchers in STEM subjects. This section will also be expanded in May.<\/p>\n

Qx Anything else you wish to add?<\/strong><\/p>\n

Alexandra Mousavizadeh:\u00a0<\/strong>We love hearing from our readers at Tortoise; the point is that we include them as part of the conversation. Our AI Sensemaker newsletter contains our and their thinking, keeping you posted on the latest developments in AI, and comes out once a fortnight. You can sign up here<\/a>. \u00a0To get more involved, you can also apply to join our AI Network<\/a>: it\u2019s a global community of experts, policy makers, and business leaders, who take part in monthly round tables. These cutting-edge conversations set the pace for all things AI.<\/p>\n

————————-<\/p>\n

\"Alexandra\"<\/a><\/p>\n

Alexandra Mousavizadeh<\/strong> is a Partner at Tortoise Media, running the Intelligence team which develops indices and data analytics. Creator of the recently released Responsibility100 Index and the new Global AI Index. She has 20 years\u2019 experience in the ratings and index business and has worked extensively across the Middle East and Africa. Previously, she directed the expansion of the Legatum Institute\u2019s flagship publication, The Prosperity Index, and all its bespoke metrics based analysis & policy design for governments. Prior roles include CEO of ARC Ratings, a global emerging markets based ratings agency; Sovereign Analyst for Moody\u2019s covering Africa; and head of Country Risk Management, EMEA, Morgan Stanley.<\/span><\/i><\/p>\n

Resources<\/strong><\/p>\n

– The Global AI Index<\/a><\/p>\n

– The Global AI Index Methodology Report<\/a><\/p>\n

– The Tortoise Global AI Summit: the Readout. Thursday 10 December 2020<\/a><\/p>\n

Related Posts<\/strong><\/p>\n

– On The Global AI Index. Interview with Alexandra Mousavizadeh. ODBMS Industry Watch.by Roberto V. Zicari on January 18, 2020<\/a><\/p>\n

Follow us on Twitter: @odbmsorg<\/a><\/strong><\/p>\n

##<\/p>\n","protected":false},"excerpt":{"rendered":"

“I think the conversation is really about a shift of where funding is coming from. Governments are spending far less than the big tech platforms. What this tells us about who owns the direction of travel of AI is fascinating. Are we now in a position where the power that the public sector was able […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[1],"tags":[990,1533,881,1639,1535,1638,1637,629],"_links":{"self":[{"href":"https:\/\/www.odbms.org\/blog\/wp-json\/wp\/v2\/posts\/5334"}],"collection":[{"href":"https:\/\/www.odbms.org\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.odbms.org\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.odbms.org\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.odbms.org\/blog\/wp-json\/wp\/v2\/comments?post=5334"}],"version-history":[{"count":20,"href":"https:\/\/www.odbms.org\/blog\/wp-json\/wp\/v2\/posts\/5334\/revisions"}],"predecessor-version":[{"id":5355,"href":"https:\/\/www.odbms.org\/blog\/wp-json\/wp\/v2\/posts\/5334\/revisions\/5355"}],"wp:attachment":[{"href":"https:\/\/www.odbms.org\/blog\/wp-json\/wp\/v2\/media?parent=5334"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.odbms.org\/blog\/wp-json\/wp\/v2\/categories?post=5334"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.odbms.org\/blog\/wp-json\/wp\/v2\/tags?post=5334"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}