AI Index 2019: UK lags behind competitors for real-world impact
The 2019 edition tracks three times as many data sets as last year's report, with the HAI producing two additional tools to help experts navigate the data
The latest AI Index 2019 Report has been published by the Stanford Institute for Human-Centered Artificial Intelligence (HAI), and while the UK cemented its position as a powerhouse for AI-focused research, statistics show that the nation currently lags behind its peers in terms of real-world impact.
The 2019 edition tracks three times as many data sets as last year’s report, with the HAI producing two accompanying tools to help navigate the data: The Global AI Vibrancy tool, which compares 28 countries’ global activities across 34 indicators – including both a cross-country perspective and country-specific drill-down; and the AI Index arXiv Monitor, a full research paper search engine that helps people conduct their own research on the topic.
It has become increasingly common for AI-focused companies to collaborate with academics on research. According to the report, these partnerships are most prevalent in the US, China, Japan, France, Germany, and the UK. US authors are generally cited 40% more than the global average.
In 2018, 37% of published AI papers were attributed to East Asia and the Pacific, 24% to Europe and Central Asia, and 22% to North America. South Asia’s share in international AI journal publications grew to almost 8% last year. In terms of average journal publications per million people between 2015-18, China had the highest volume, followed by the US, India, the UK, and Germany
To provide a detailed overview of AI skill penetration across sectors and countries, the report used a sample of the top five industries with the highest global AI skill penetration, including software and IT services; hardware and networking; education; finance; and manufacturing. The US ranks among the top five across all sectors, while China only lands a top five ranking for education-related skill penetration. India, the US, France, China, and Israel are frequently among the top countries.
Other pockets of specialisation by country include Norway and Israel for AI skills in software and IT; Norway, France, and Sweden in hardware and networking; France, Israel, and Sweden in hardware and networking as well as manufacturing; Spain and Switzerland in education; and the UK and Canada in Finance.
In related news: Most teachers fear AI’s impact in the classroom
Countries differ considerably with respect to which AI ethical issues received the most news coverage. While media platforms in the US and UK demonstrated fairly balanced coverage between categories, others reflected more specific focus areas. In Switzerland, for example, 45% of all articles covered guidelines and frameworks on AI development, while 44% of AI news in China was about safety and security, and 48% of articles in Singapore discussed transparency and ‘explainability’.
In terms of the UK government’s perception of the topic, mentions of AI and machine learning in parliamentary proceedings were scarce prior to 2016. Mentions peaked in 2018, declining in 2019 despite the subject remaining significant.
“AI needs to get out of the lab to boost Britain to new heights,” said Dr Iain Brown, head of data science at SAS for the UK and Ireland.
“In 2020, the transformative power of AI must be more than a figment of our imaginations. With R&D firmly underway, it’s time to let AI produce tangible differences to workplaces, maximising productivity and efficiency, and liberating workers from mundane tasks. Deploying AI will give British businesses fresh insights that could not only solve stagnant global economic growth, but uplift global GDP by US$15 trillion in the next 10 years.
“In this uncertain time for Britain, being at the forefront of business innovation will ensure that British businesses leverage the seismic change AI is set to bring in the coming decade,” added Brown.