“As the pace of breakthroughs in artificial intelligence research slows, companies like Alphabet and Microsoft are reallocating researchers to products that generate revenue and away from pure AI research. This shift suggests that the market for artificial intelligence hiring is entering a new phase,” reports Kevin McLaughlin for tech news site, TheInformation.com.
Patents
From 2012 to 2017 AI patent families grew faster than AI scientific publications. The ratio of scientific papers to patents has fallen dramatically as machine learning finds a greater number of commercial applications. For instance, cloud computation has been monetized but not AI hosted services. Computer vision has been the most popular area for patents and within computer vision, biometrics applications tops the list.
Publishing
In the publishing world, Europe publishes the most AI papers but only China’s average citation rate has been growing. Europe publishes 16,000 AI papers annually, China 15,000, U.S. 10,000 and the rest of the world 18,000. The U.S. citation impact has remained the same at 1.8 field weighted, while Europe’s has been going down at 1.1 and China at .9 is increasing.
Academic
China continues to lead the way in university AI course enrollment with 16x more students enrolling in AI course at Tsinghua University since 2012 with Stanford only having 5x higher enrollment. In a study done on papers published at 21 machine learning conferences by 22,400 unique authors, 44% of the authors earned their PhD in the U.S., 11% in China and 6% in the U.K. Five countries accounted for the employment of 72% of these authors with the U.S. leading the way. Academic enrollment is important because as artificial intelligence hiring increases, as it did with Microsoft AI and Research Group going from 5000 people in 2016 to 8000 in 2017, companies are hiring academics in part to get access to their students for hiring when they graduate.
Interestingly the U.S. and China artificial intelligence hiring markets are more mature with low inflows and outflows of talent. Canada, the U.K. and Switzerland are termed “platform countries” because they both attract foreign talent and export locally trained talent. The good news is American Universities are stepping up to the artificial intelligence hiring need. A recent $1 billion investment in computing and AI at MIT is allowing for 50 new faculty positions doubling MIT’s academic capability in this field. The University of Virginia also recently announced a $130 million donation to create a dedicated data science school.
Compensation
Compensation of senior engineers at large tech companies is approaching $1 million according to the 2019 State of AI report.. Base salaries are just under $300K with bonus’ over $100k but stock is where the real impact is with valuations as high as $400K. At the other end of the spectrum is the huge growth in data labeling jobs that pay 10 Yuan or $1.47/hour. Beijing based Mada Code has Microsoft and Carnegie Mellon as customers and claims to have 20,000 freelancers working for them labeling data.