The five most important new jobs in AI, according to KPMG

AI’s salad days.
AI’s salad days.
Image: AP Photo/Eric Risberg
We may earn a commission from links on this page.

Perhaps as a counter to the panic that artificial intelligence will destroy jobs, consulting firm KPMG today published a list of what it predicts will soon become the five most sought-after AI roles. The predictions are based on the company’s own projects and those on which it advises. They are:

  • AI Architect – Responsible for working out where AI can help a business, measuring performance and—crucially— “sustaining the AI model over time.” Lack of architects “is a big reason why companies cannot successfully sustain AI initiatives,” KPMG notes. 
  • AI Product Manager – Liaises between teams, making sure ideas can be implemented, especially at scale. Works closely with architects, and with human resources departments to make sure humans and machines can all work effectively.
  • Data Scientist – Manages the huge amounts of available data and designs algorithms to make it meaningful.
  • AI Technology Software Engineer – “One of the biggest problems facing businesses is getting AI from pilot phase to scalable deployment,” KPMG writes. Software engineers need to be able both to build scalable technology and understand how AI actually works.
  • AI Ethicist – AI presents a host of ethical challenges which will continue to unfold as the technology develops. Creating guidelines and ensuring they’re upheld will increasingly become a full-time job. 

While it’s all very well to list the jobs people should be training and hiring for, it’s another matter to actually create a pipeline of people ready to enter those roles. Brad Fisher, KPMG’s US lead on data and analytics and the lead author of the predictions, tells Quartz there aren’t enough people getting ready for these roles.

Training in AI is increasing “fairly rapidly” at universities and other institutions, but not rapidly enough, Fisher says, meaning that demand will continue to outstrip supply for some time. While some roles could conceivably be filled by existing employees—that of ethicist, for example—others need specific technical know-how that might best be gained through external training.

“The bottom line is that a minority of these skills may already exist in most organizations; however, the vast majority simply don’t exist,” Fisher says. “These positions will need to be filled by a combination of retooling/retraining existing resources or hiring new resources altogether.”

The gap presents an opportunity both for those entering the workforce and those looking to move within it. Fisher has a steer for those who are eyeing AI jobs but have yet to choose an academic path: business process skills can be “trained,” he said, but “there is no substitute for the deep technical skillsets, such as mathematics, econometrics, or computer science, which would prepare someone to be a data scientist or a big-data software engineer.”

Someone with “no technical skills probably cannot be a data scientist, but they might have the background to be an AI ethicist or even an AI project manager,” he advises. “While it’s unlikely they would have the right skills “out of the box”, with some degree of effort they can be the AI specialist of the future.”

For those who are already working but want to move into an AI role, Fisher says they should choose the type of role that best suits their existing skills, and prepare to “retool;” he notes that progressive companies are already helping employees to do just this, in anticipation of a big shift in skills needs. 

Joanne Cash, chair of the board of Mind Gym, which uses behavioral science to create bespoke training, says that while some jobs will indeed disappear as the use of AI and automation grows, the need for other kinds of jobs with nuanced skills will increase.

“You’ll see a reduction in the number of young financiers needed, for example, because there are a large number of financials which can be automated,” she says, “but an increasing number of highly skilled managers [will be] needed to look after those who are left in roles.” she says.

As new jobs are created and others lost, the demarcation between human and machine is likely to become clearer rather than more blurred: The new roles will be based on “the ability to be compassionate, empathetic, to have emotional intelligence,” Cash says.