Stanford Report Distills the Global Impact of AI and Machine Learning

Did you know that the time required to train a large image classification system on cloud infrastructure has fallen from about three hours in October 2017 to about 88 seconds in July 2019? Or that AI has become the most popular specialization among computer science Ph.D. students in North America? Or that the U.S. share of total jobs in AI-related fields posted in 2010 grew from 0.26 percent in 2010 to 1.32 percent in 2019?

These are just a few of the findings published in Stanford University's "AI Index 2019 Report." An independent initiative within the university's Human-Centered Artificial Intelligence Institute (HAI), the report is the result of a collaborative effort led by the AI Index Steering Committee, which is an interdisciplinary group of experts from both academia and the enterprise, and more than 35 sponsoring partners and data contributors.

The HAI's mission is to "advance AI research, education, policy, and practice to improve the human condition," its Web site states. It focuses on studying, guiding, and developing human-centered AI technologies and applications. "We believe AI should be collaborative, augmentative, and enhancing to human productivity and quality of life," the site states.

This was the HAI's third edition of the annual report. Each report has expanded the coverage of their predecessors. The 2019 report, published in mid-December, tripled the data sets of the 2018 report. In fact, there's so much data in this report, The HAI created the Global AI Vibrancy Tool, which compares the global activities of the different countries covered. It includes both a cross-country perspective and an intra-country drill down.

"It is tempting to provide a single ranking of countries," the report's authors state on the Web site, "but such rankings are notoriously tricky. Projecting complex, heterogeneous measures down to a single number (or even a small set of numbers) is fraught with methodological subtleties, and can be highly subjective or skewed. For now, we thought it more responsible and useful to provide a tool for the reader to set the parameters and obtain the perspective they find most relevant."

The report organizes all that data into nine subject areas (chapters), including research and development, conferences, technical performance, the economy, education, autonomous systems, public perception, societal considerations, and national strategies and global AI vibrancy. The Research and Development chapter, for example, examines bibliometrics data, including volume of journal, conference, and patent publications, as well as their "citation impacts" by world regions. For good measure, the chapter includes Github Stars for key AI software libraries to provide a measure of the popularity of various AI-programming frameworks. It includes a graphs showing the number of times various AI and ML software packages have been starred on GitHub.

That chapter also covers the gender diversity of AI researchers based on the arXiv (pronounced "archive") preprint repository of scientific papers not yet peer reviewed, but approved for posting after moderation. The researchers developed the ArXiv Monitor full-paper search engine to automatically and continuously track technical metrics from papers published on arXiv.

The Stanford researchers were also especially cognizant of the potential for bias to skew results. "Given that measurement and evaluation in complex domains remain fraught with subtleties," they said in a statement, "the AI Index has worked hard to avoid bias and seek input from many communities." They even held a workshop at Stanford ("Measurement in AI Policy: Opportunities and Challenges") prior to publication that brought together more than 150 industry and academic experts from a variety of disciplines related to AI to discussed "the many pressing issues that arise from data measurement of AI."

In just three years, the Stanford "AI Index 2019 Report" has become an essential resource for anyone working with or thinking about AI/ML technologies -- which, nowadays, is just about everybody.

About the Author

John K. Waters is the editor in chief of a number of sites, with a focus on high-end development, AI and future tech. He's been writing about cutting-edge technologies and culture of Silicon Valley for more than two decades, and he's written more than a dozen books. He also co-scripted the documentary film Silicon Valley: A 100 Year Renaissance, which aired on PBS.  He can be reached at