National AI Policy Making — Part 2: Review and Comparison of National AI Policies

6 min readDec 24, 2018

Part 1 of this series of posts about national AI policy making presented an overview of the challenges and opportunities for nations. According to Gartner Inc., AI is expected to generate a total of $1.2 trillion in global business value in 2018, with AI-derived business value forecast to reach $3.9 trillion by 2022. However realizing these benefits is not simple.

“AI is a broad field that involves extremely disparate disciplines. It’s difficult to have state-of-the art expertise in all of those aspects in one place.” — Carnegie Mellon University website.

In this post I’d like to provide an overview of the national AI policies of a few select nations so that you can get an idea of their views, approach and preferences. My goal is to present elements of policies from a diverse set of nations (geographical regions, demographics, economic situation) but I acknowledge that it is difficult to provide a comprehensive treatment.

Timeline of AI Policies — Via Oxford Institute

The goal of any technical policy is to use science and engineering to inform intelligent, responsible strategies and policies to benefit communities from local to global (adapted from Technology and Policy Program, MIT.)

Obviously different countries want to use their strengths to gain advantages through AI. For instance certain countries have strong research centers, intellectual property or industrial knowledge, others have a huge base of educated youth and others have piles of cash to throw at new technologies.

Here’s how I did my homework for this post.

Reviewed the policy documents of various countries — used a list of criteria (see below for more) to compare them
Online search led me to work done by other researchers who compiled AI policies of different countries and prepared useful illustrations such as timelines, feature summary and more (the image below is from this source.) I also looked at the work by policy think tanks and policy focused groups at selected universities (CMU, MIT, Stanford).
AI Manifestos and declarations from top companies that are leading work on AI — e.g. Google’s AI principles (socially beneficial, avoid bias, safety, accountability, scientific excellence, limit harmless use). Microsoft and IBM have similar declarations. Governments can certainly use these.
Researched structured, data-driven approaches to how AI efforts can be measured — example: AI Index, an effort to track and collate AI data and analysis for policymakers, researchers and others.

How do you compare something as broad as a national policy? Here’s my list of criteria (see part 1 of this series for the 7 areas in AI policy):

Vision — coherence, clarity, ambition, scope, priorities, focus areas
Quality of Plan — completeness, digital maturity, metrics, realistic goals, metrics
Resources — budget, leadership, collaboration and partnerships
Governance Model — management structure, resource allocation, regulation and legislation

Based on the factors above one could derive overall rank of each policy. The AI Index work is a good source but lacks data for most countries. Oxford Institute has published their list of ‘Government AI readiness index’ of OECD countries — below is a snapshot of their approach and data sources so you can make your own judgements about the reliability of this approach. I would caution against reading too much into their ranking system.

Oxford Institute — AI Readiness for Governments

Below is a summary of my reading with links to additional information for National Policy for selected countries and regions — this information could be official document and/or article with a summary and additional links to relevant studies.

North America:

US — The leader in research, business and defense AI (competing with China), home to leading software and hardware companies and top talent; current government has no published policy — free market approach with minimal regulation.
Canada — First to publish a policy, strong focus on research and talent, CIFAR leading the effort with 3 new AI institutes; leadership includes top AI minds of today

Europe:

EU — The European Union published its strategy for AI in April 2018, noting EU’s strengths and future plans — here’s a nice summary.
France — Announced a comprehensive policy (235 pages) with a sharp focus on humanity. The president of France has called it top priority. This report is worth a read - here’s a good blog summary and here’s the 10-page official brief in English.
UK — No official policy but a sector deal that sets out actions to promote the adoption and use of AI in the UK. It includes tax credits for R&D, changes in education policies and more.

Asia:

China — Issued guidelines on AI. China leads the world in many aspects of AI, especially sophisticated collection and use of data and active role played by government and rollout of AI technology to public and commerce applications. Needs more fundamental research and open environment.
Japan — Second largest AI country in Asia after China and a top producer of advanced robots. Japan is pursuing comprehensive AI policies and plans with a focus on industrialization.
India — Policy focused on inclusion and building on its education system, data-generating consumers, IT workforce and research centers. Lack of structure and coordination of data and lack of coordination between development agencies threaten to slow its progress.
Singapore — Included here because of its distinct focus and efforts that can guide others — more info in this post by Tim Dutton
UAE — Appointed first Minister of State for AI. Released vision ‘AI 2031’ with a marketing video — UAE has a track record of early technology adoption, lets see how the AI ambitions pans out.

As you may have noticed I included more countries from Asia. With its large, educated and younger population, Asia is poised to become a frontrunner in the development and adoption of AI. China is already the leader in many aspects of AI. More government intervention and support is needed to give Asian private sector a chance to help broader national objectives.