What Salary Premiums Will Employers Pay for AI-Skilled Workers?

If American companies really are in the middle of an artificial intelligence recruiting frenzy, as the Wall Street Journal first reported in August 2023, one would expect to observe plenty of evidence that hot demand is bidding up wages for employees with AI skills. The Journal article titled “The $900,000 AI Job Is Here” reports plenty of job listings and anecdotal evidence supporting such a trend, especially for jobs requiring data science and machine learning experience in working with AI systems. However, survey research data on higher wages paid to AI-skilled workers has been hard to find so far.

Just now we’re witnessing new survey data indicating how much more money workers should be able to earn as soon as they can demonstrate “upskilling” through training and experience with artificial intelligence platforms. The new data showed up in November 2023, only a few days before the one-year anniversary of the explosive launch of the ChatGPT artificial intelligence platform back on November 30, 2022.

This new wage data appears within the results of an employer survey released by Amazon Web Services that’s entitled “Accelerating Al Skills: Preparing the Workforce for Jobs of the Future.” The AWS results build upon a slightly earlier survey by edX that contrasted opinions about artificial intelligence by CEOs and chief executives with views about AI by lower-level staff. We covered this edX survey in a November 2023 feature article on our sister site BSchools entitled “Analysis: Half of CEOs Say AI Could Replace Most of Their Role, Survey Says.”

However, although the edX poll provided some of the first indications that American companies were experiencing substantial challenges in recruiting enough workers with AI skills—a phenomenon now known as the AI talent gap—edX stopped short of asking questions about the additional salary value of those skills.

For the AWS survey, online fieldwork took place during August and September 2023 by the London-based tech polling and management advisory firm Access Partnership. A fairly robust representative sample of 4,637 respondents participated across the United States, composed of 1,340 employers and 3,297 employees between 18 and 74 years old. Pollsters restricted the sample to organizations and workers who utilize at least some tech skills, ranging from the basics like word processing to advanced capabilities like software development.

The AWS survey asked respondents for their views related to three general themes:

• The current development of and future needs for tech skills
• The perceived benefits of artificial intelligence, including generative Al
• The current and future state of training for Al skills, including barriers to obtaining training in those skills

Amazon’s survey disclosed that about three-quarters of employers who’ve prioritized hiring employees with AI skills (73 percent) say they’re having trouble finding candidates who are qualified. The poll also revealed that more than two in five employers (42 percent) are actively seeking employees qualified in AI development. What’s more, 93 percent of the respondents work for firms that expect to apply AI-based technologies by 2028.

Dramatic poll results like those suggesting hot demand for workers with AI skills seem more credible when one considers some recent news about artificial intelligence. For example, a widely reported September 2023 analysis from the management consulting firm McKinsey & Company predicts that AI could add trillions of dollars worth of value to the worldwide economy by automating up to 70 percent of business functions.

Another study by IDC that Microsoft sponsored had reported in November 2023 that companies realize $3.50 in returns on average for every dollar they invest in AI. In that poll of 2,000 business leaders from around the world, 52 percent of the respondents reported that their toughest obstacle in implementing the technology is a shortage of workers with AI skills. (However, keep in mind the potential for bias in these results, given that Microsoft had already invested $13 billion in OpenAI, the San Francisco research company that develops ChatGPT.)

Now, back to the Amazon survey. The AWS pollsters asked employers what sort of a wage premium they’d be willing to pay to hire an employee who had invested the time and effort needed to acquire artificial intelligence skills. This kind of premium goes by various names such as “salary boost” or “salary uplift,” and as we also point out over on BSchools, it provides a widely-accepted method of comparing the compensation benefits resulting from various forms of education and training—such as online MBA degree programs.

Believe it or not, employers told the AWS pollsters that they’d be willing to pay workers who acquire AI skills a minimum wage premium of at least 30 percent, irrespective of the employee’s function within the organization.

Information technology workers would see the greatest gains at 47 percent of their current salaries, probably reflecting the remarkable and widely reported ability of generative AI platforms like GitHub Copilot to rapidly write and debug software code. However, employers claim that they’ll reward workers across the organization with a pay boost if they acquire AI skills. Here’s how the employers reported that such a weighted-average, percentage salary boost would vary by department:

Compared to typical annual raises for most professionals, these wage increases are substantial. How do employers justify these large salary boosts? Here’s what the pollsters report:

The anticipated pay premiums across departments [are] because Al’s key benefits—automating tasks, boosting creativity, and improving outcomes—have dispersed applications across departments and tasks. Employers anticipate that workers with Al skills will be able to drive additional productivity and higher-quality work, which would command a salary increase.

Digging into the data, one quickly notices a pattern: the salary boosts roughly correlate with the productivity boosts that employers expect from their workers who use artificial intelligence platforms. For example, employers believe that the average productivity boost amounts to about 47 percent, but the gains vary from 43 percent to 49 percent depending on company size.

Enhancing technical and coding skills through techniques such as prompt engineering—a skill we discuss below—appears to be only one part of the artificial intelligence skills challenge. According to the pollsters, the implications for workers are clear: “Taking advantage of Al is going to require a far broader set of skills than just coding.”

The data shows that the employers have identified a set of foundation skills that they require for their employees to use AI well. It turns out that these are classic skills typically taught within most of the better undergraduate and graduate degree programs, whether on-campus or online.

For example, the employers actually cited two skill groups as more important than technical skills among employees who use artificial intelligence platforms well. The first group includes critical thinking and problem solving—which 61 percent of employers ranked among their top three foundational skills—and the second group involves creative thinking. “Critical thinking is essential to evaluate the accuracy and relevance of Al outputs, while problem-solving helps optimize the capabilities of Al systems by defining and structuring analyses appropriately on available data,” write the pollsters.

Interestingly enough, the employers ranked a surprising class of capabilities just below technical skills: ethics and risk management. Here’s how the pollsters explain this unexpected choice:

Ethics and risk management is also ranked as the fourth most important skill needed to use Al effectively. That’s because while Al can mimic many human skills and competencies, it still falls short in other areas, like emotional intelligence, contextual understanding, common sense, adaptability, ethics, and intuition. Ethics and risk management is therefore critical to guard against potential Al risks, including algorithmic bias, lack of accountability, compliance challenges, and the need for transparency and safety measures.

Along with releasing the survey results, Amazon also announced a new learning initiative called AI Ready, which aims to provide two million people worldwide with free AI training by 2025. The company claims that it intends this initiative to open opportunities aligned with in-demand jobs for those currently in the workforce and college and high school students.

“Artificial intelligence is the most transformative technology of our generation. If we are going to unlock the full potential of AI to tackle the world’s most challenging problems, we need to make AI education accessible to anyone with a desire to learn,” said Swami Sivasubramanian, the vice president of data and AI at Amazon Web Services. “The goal of AI Ready is to help level the playing field of AI education, supported by the new initiatives we’re launching today. We will also scale our existing free AI training programs and courses as we continue to remove cost as a barrier to accessing these critical skills.”

The initiative offers eight new free courses covering various aspects of generative AI that we briefly profile below. However, we don’t agree with Amazon’s classification of some of the course topics. Amazon implies that some of these courses are only intended for software developers and technical audiences, even though the courses teach skills essential for business and nontechnical audiences as well.

For example, we believe that their prompt engineering course teaches a basic generative AI skill that everyone needs to learn. We take an in-depth look at this skill in a May 2023 feature article here on OnlineEducation.com titled “The $335,000 ChatGPT Skill Savvy Online Students Need to Know.”

Introduction to Generative Artificial Intelligence delivers an introduction to generative AI along with essential concepts, such as foundation models.

Generative AI Learning Plan for Decision Makers shows students how to plan a generative AI project and “build a generative AI-ready organization.”

Introduction to Amazon CodeWhisperer shows participants how to produce lines of code using Amazon’s code generator.

Foundations of Prompt Engineering teaches students the basics of designing input language for AI tools, followed by advanced techniques.

Low-Code Machine Learning on AWS explores how to prepare data, then train and deploy machine learning models with minimal coding.

Building Language Models on AWS shows how to build language models and tune foundation and open source models.

Amazon Transcribe—Getting Started covers how to convert speech to text using this managed AI service.

Building Generative AI Applications Using Amazon Bedrock teaches how to build AI applications using the company’s Bedrock platform.

Amazon also pledges to pay for a certificate course from Udacity for 50,000 college and high school students from underrepresented and underserved backgrounds. To pay Udacity’s tuition, AWS has set aside a $12 million scholarship fund.

The course, Introducing Generative AI with AWS, explores foundational concepts in generative AI and guides students through a project. Students who complete the course earn a certificate from Udacity so they can “showcase their knowledge to future employers.”