Ranking governments on AI - it's time to act
By Isak Nti Asare and Hannah Miller
Map displaying the results of Oxford Insights' 2017 Government AI Readiness Index
This week, the Economist Intelligence Unit, sponsored by the tech firm ABB, released their report The Automation Readiness Index: Who is Ready for the Coming Wave of Automation? The paper quickly attracted sensationalist headlines about the ‘rise of the robots’, similar to some of those we saw last year when we released our Government AI Readiness Index.
Although at first glance the two indexes seem to be measuring similar things, a look at the EIU’s report and input metrics reveals a number of interesting differences. The principal difference is what each index sets out to measure. Our index specifically set out to answer the question: how well placed are the national governments in the OECD to take advantage of the benefits of automation in their operations? In other words, our index captures the capacity of national governments to use AI to help transform their own public service delivery and operations.
The EIU’s index has a different focus. Their index sets out to “determine which countries are better positioned to take up the policy challenges that automation poses”. The inputs focus on three areas: the innovation environment that helps support research into AI and AI-related businesses; education policies that help to develop the human skills needed to develop and operate these technologies; and labour market policies designed to manage the transition to an automated economy. The EIU’s index therefore focuses on capturing certain aspects of the policy environment of a given country, as a composite indicator of how prepared that country is for the implementation of AI on a grand scale in society.
We believe strongly in the importance of developing high-quality, in-country AI talent if a country is to thrive in the so-called fourth industrial revolution. There was a commendable emphasis in the EIU’s report on education policies, with indicators ranging from technology education programmes, to career guidance programmes, to the use of AI and data in education. Implementation of AI requires a much broader perspective, which the EIU sought to capture through 52 indicators, divided into the focus areas discussed above. Our more narrowly focused index is comprised of nine input metrics, divided into three subgroups (public service reform, economy and skills, and digital infrastructure). The EIU made interesting use of qualitative inputs, such as interviews with experts from around the world, to add some context and colour to their report.
The EIU posits that robust policy responses to address the challenges and opportunities of AI have so far been lacking. We disagree. Governments have already begun to create wide ranging and forward-thinking policy documents to lay out strategies for how they plan to maximise AI’s benefits, and mitigate its potentially negative impacts. France’s strategy for example, #FranceIA, examines AI’s likely social and economic impacts, and sets out to prepare the country for the future through training and research. There is a clear emphasis in the strategy of creating a multi-stakeholder approach. We recently published a blog reviewing existing national AI strategies, and the five key themes that these strategies share:
- Using AI in government and public services;
- Research and development;
- Capacity, skills and education;
- Data and digital infrastructure.
Since then, new countries have produced comprehensive strategies for AI, including Mexico, whose forthcoming AI strategy is to be based on work carried out by Oxford Insights in collaboration with our Mexican partners C Minds. Other governments, such as India, are currently working to develop theirs. 25 European countries recently signed up to the EU’s Declaration of Cooperation on AI, pledging to work together to maximise the benefits and minimise the challenges of automation.
Both indexes illustrate a stark reality: governments cannot afford to avoid the rise of AI. Those who fail to take action will be left behind. Our Government AI Readiness Index shows that many national governments are ready to proactively engage with these debates. The EIU’s implementation index complements ours, in showing why it is important that governments take action now. The disparity between capacity and implementation is seemingly large, but we must not discount the agency of governments to create an environment to take advantage of the benefits of AI, while mitigating its challenges. It is encouraging that many governments have already begun to act.
Oxford Insights will be publishing our annual Government AI Readiness Index later in 2018. If you have suggestions for this year’s Index or on our methodology, please get in touch at firstname.lastname@example.org. We also welcome interest from organisations looking to sponsor our research.