The Artificial Intelligence Index Report 2019 , the independent annual report published by the Human-Centered Artificial Intelligence Institute at Stanford University, which presents the results of advances in research on artificial intelligence, is now published . This edition has collected three times more data than that of 2018 and is accompanied by the Global AI Vibrancy Tool, a tool that compares the activity of all countries individually.
The report is public and can be read in PDF from this link . In this article we are going to highlight some of the most relevant points and, after learning about them, we will talk about the Spanish case to find out what our country’s position is in the artificial intelligence sector.
Growth of articles related to artificial intelligence published in all publications between 1998 and 2018.
We start talking about research and development, and according to the AI Index Report of 2019, the volume of articles on artificial intelligence peer-reviewed has grown by 300% in 2018 compared to 1998 . China continues to be the country that publishes the most articles on artificial intelligence, to the point that it has surpassed Europe as a whole. However, the United States maintains the leadership in terms of relevance (which is measured according to the impact of the citations that its works receive), with a Field-Weighted Citation Impact 50% higher than that of the Asian giant.
Another aspect that highlights the importance of the United States in the artificial intelligence sector is that North America represents more than 60% of the world activity of AI patent citations between 2014 and 2018. East Asia is the protagonist of 32% of the citations in artificial intelligence magazines worldwide. As relevant data, women have a “relatively high presence” in AI research in Western Europe, mainly in the Netherlands and Denmark, as well as in Argentina, Canada and Iran.
Growth of publications on artificial intelligence in China (blue), United States (gray) and Europe (dotted line).
The growing interest in artificial intelligence can also be seen in the number of people attending conferences. A good scale is NeurIPS, the largest in the world, which expects the 2019 edition (which is taking place these days) to exceed 13,500 attendees , 41% more than in 2018 and 800% compared to 2013. Others Conferences such as AAAI and CVPR have also experienced a greater influx with a 30% growth in attendees.
Further research, the report notes, results in faster and cheaper algorithms to train
Women in Artificial Intelligence (WiML) move along the same lines, which has had eight times more participants than in 2014 and AI4ALL, which has 20 times more alumni than in 2015. For the AI Index Report 2019, “these increases reflect an effort continued to include women and underrepresented groups in the field of AI. ”
There has also been a growing interest in the field of education, since “enrollments continue to grow rapidly in subjects related to artificial intelligence, both in traditional US and international universities, as well as in the online offer” . In fact, academically, AI has become the most popular major among Computer Science PhDs in North America, with the number of students having doubled. In 2018 alone, 21% of these PhDs specialized in AI and machine learning.
The industry has become the main talent tractor. In 2004, the industry had 20% of doctorates in artificial intelligence, while in 2018 the percentage has risen to 60% . Regarding gender, “the diversification of AI teaching staff based on gender has not shown great progress, as women represent less than 20% of the new professors hired in 2018. Similarly, the proportion of women with PhDs in AI has remained virtually constant at 20% since 2010 in the United States. ”
Growth in job openings related to artificial intelligence in the United States since 2010.
Advances in artificial intelligence are also helping to create jobs. Singapore, Brazil, Australia, Canada and India are the countries with the highest growth in hiring for jobs related to artificial intelligence since 2015. The case of the United States is striking, whose job offers related to AI have risen to 1.32 % of total job offers in 2019 . In 2010 they were 0.26%. Most of the positions are related to machine learning and, mainly, they are related to the manufacturing and high-tech services sector.
Investment has also grown significantly. Investment in start-ups related to artificial intelligence maintains “sustained growth”, so much so that if 1.3 billion dollars were invested in 2010 in 2018 the figure has risen to 40.4 billion . That means that the growth rate in investment is 48% year-on-year. A lot of money. The total investment in 2019 has been 70,000 million dollars.
Autonomous cars have been the main beneficiaries of the investment in 2019, remaining with 7,700 million dollars or what is the same, 9.9% of the total . It is followed by research in drugs, cancer and therapies ($ 4.7 billion), facial recognition (another $ 4.7 billion), video content ($ 3.6 billion) and bank fraud detection ($ 3.1 billion) .
Now that we are talking about autonomous cars, an interesting figure is that the total number of miles traveled by autonomous cars in the State of California has risen to two million miles, 3.21 million kilometers , in 2018. There are currently more than 500 cars freelancers from more than 50 companies walking the streets, which means that they have multiplied by seven between 2015 and 2018.
Finally, the report addresses the perception of society and existing concerns regarding the advancement of artificial intelligence. On the one hand, the report states, “the communications of the world’s central banks demonstrate a great interest in artificial intelligence , especially on the part of the Bank of England, the Bank of Japan and the Federal Reserve of the United States.”
This growing interest from the world’s major financial institutions has also led to a “significant increase in AI-related legislation in the records of (the United States) Congress, in committee reports, and in legislative transcripts from around the world. “. It seems, therefore, that countries are beginning to introduce the regulation of artificial intelligence in their agendas , although approvals of specific laws or regulations are not mentioned.
Main concerns of the documents on “Principles of Artificial Intelligence”.
Regarding concerns, the report has analyzed 59 documents on principles of ethics in artificial intelligence and the result is clear: the most important challenges are impartiality, interpretability and explicability and transparency . It coincides in a way with the data obtained after analyzing 3,600 press articles worldwide, whose concerns revolve around the legislative frameworks of AI, data privacy, the use of facial recognition, algorithm bias and the role of big technology.
Finally, the report highlights that artificial intelligence can contribute to each and every one of the 17 United Nations Sustainable Development Goals through “identified use cases.” Above all, the report highlights up to 29 cases in the objective of well-being, 24 cases in peace, justice and strengthening of institutions, 13 cases in educational quality and 10 in the achievement of non-poverty. However, there is still a need to “overcome bottlenecks to deploy AI for sustainable development at scale.”
As we said, AI Index Report comes accompanied by a report broken down by country and Spain is among them. It is difficult to measure the performance of countries in terms of artificial intelligence since, as explained on the Vibrancy Index website, “projecting complex and heterogeneous measures down to a single number (or even a small set of numbers) is fraught with methodological subtleties and it can be very subjective or biased. ” Simply put, they are guidance metrics.
To measure the performance of a country, an average of 33 points is made, which are divided into three categories: research and development, economy and inclusion . For example, R&D is scored by the number of AI patents or the number of journal articles on AI. In economics, enrollments in AI and private investment in AI are considered, and in inclusion the percentage of female authors and the number of AI-related positions held by women.
How is Spain? Below average relative to high-income countries . We lag behind somewhat in research and development and economics, although we are above average in inclusion. If we look at the complete photograph, Spain is somewhat below the average of the rest of the countries considered, standing out mainly in the percentage of enrolled in subjects of artificial intelligence, percentage of hires and proportion of female authors. However, the average is reduced because Spain does not stand out in cited works on artificial intelligence, patents, per capita patents or private investment, among others.