The short answer: an astronomical sum.
The New York Times recently analyzed a tax filing by OpenAI, a nonprofit that specializes in A.I. work, and found that its top researcher, Ilya Sutskever, made nearly $2 million in 2016. Another researcher earned $800,000.
“I turned down offers for multiple times the dollar amount I accepted at OpenAI,” Sutskever told the newspaper. “Others did the same.” His reasoning for rejecting that extra cash? He believes strongly in OpenAI’s mission to help create A.I. that’s friendly to humanity; in pursuit of that goal, the organization not only conducts research and publishes papers on a regular basis, but also offers a variety of tools for building and refining A.I. platforms.
Previous reports back up Sutskever’s claim that tech firms are paying out enormous salaries to those with A.I. knowledge. At a conference last summer, for example, Tom Eck, CTO of industry platforms at IBM, said that “top-tier A.I. researchers are earning the same salaries as NFL quarterbacks.” As you can imagine, many quarterbacks make many millions of dollars per year.
The Dice Salary Calculator, meanwhile, suggests that a tech pro in San Francisco with at least five years’ experience in technologies related to artificial intelligence (including data analytics and natural-language processing) can earn as much as $121,000 per year; in a “cooler” market like Kansas City, salaries can still approach (and often exceed) six figures. Maybe that’s not what an NFL quarterback makes, but that’s still really good. Take a look:
There’s a simple reason for this cash flood: tech firms believe that whoever “wins” A.I. will “win” the future (whatever it looks like), and nobody wants to lose out. But demand far outstrips supply; a recent analyst report from McKinsey & Company places the number of qualified A.I. experts in the world at fewer than 10,000. That report also suggested that practical applications of artificial intelligence could add a staggering $3.5 trillion to $5.8 trillion in value across various industries over the next several years.
For those interested in participating in this burgeoning market, there’s some good news: companies and educational organizations are rolling out more and more tools you can use to learn at least the rudiments of A.I. and machine learning. For example, Google offers a three-hour course on deep learning and machine-learning tools via its Google Cloud Platform Website, and Facebook hosts a series of videos that break down fundamental A.I. concepts such as algorithms.
And for those who want to go the MOOC route, there are Coursera and Udacity, which offer instruction in A.I., and the opportunity to self-teach via materials on GitHub. Whichever pathway you choose, you don’t necessarily need to become a top A.I. researcher in order to benefit from the opportunities in this field; even having some basic machine-learning skills can make you exponentially more attractive to a range of employers.