The entire concept of a career is being redefined as most workers move from company to company performing projects of varying lengths but in future they could be managed, at least in part, by artificial intelligence.

A new study by the human resources services and staffing company Randstad predicts that by 2019, as much as 50 per cent of the US workforce will be "agile workers", defined as people working in a temporary, contract, consultant or freelance capacity.

Already the demand for this type of worker is growing. In a 2012 Randstad survey, 18 per cent of companies said they were committed to building an "agile workforce." In this year's study, that figure jumped to 46 per cent.

"This preference for working in what we're now calling this agile model, it continues to gain steam, gain strength and gain preference with all classifications of employees, though predominantly among young workers," said Jim Link, chief HR officer for Randstad North America.

"Based on our own work experience in our own company and with our clients and customers, we expected the numbers to be up. But I think the rate that some folks responded affirmatively to the survey surprised even us."

Work experiences

So why is this happening? Clearly technology is part of it - workers are becoming more agile simply because they can. As people learn to work from anywhere, the concept of a job housed in an office becomes less central to our idea of what a career entails.

Also, as Link pointed out: "This year was the first time in our surveying where the more preferred benefit provided to employees was workplace flexibility. For years and years and years it has been health care and benefits or some version of that."

The predicted shift is away from a career in the traditional sense and toward an ongoing series of "work experiences".

A worker with a particular skill set is hired by a company to help with a short-term project. The company attracts the worker not just with pay but with a work experience that will enhance the worker's skills and marketability. The company gets the best person for the work that needs to be done, the worker gets paid and gets a valuable experience, then moves on to other opportunities.

"You can go from an experience to an experience to an experience and never work for a company, per se," Link said. "We think work is going to shift more and more that way as this whole idea of agility continues to be a collection of experiences that an employee will gather throughout the course of their lives."

Low bar

I would argue that workers attached to a company for only a short time will be less concerned with building relationships with managers, opening the door for management via artificial intelligence. (Feel free to insert your own "Seems like my manager is already artificially intelligent!" joke here.)

A recent Harvard Business Review article made the case for AI management, noting that bad human managers are so prevalent – and cause so many problems, like low morale and employee disengagement – that "the bar is quite low."

Per the article: "It would not require too much for AI to outperform average managers, let alone bad ones. It is as if the automation of drivers would have to outperform a majority of inept drivers, who crash and cause injuries to themselves and others on a regular basis."

The authors, business psychology experts from University College London, examined computer-driven management and noted advantages like more direct feedback (devoid of human bias) and better decision-making since the "human brain is incapable of processing the vast sea of data" available today and translating that information into knowledge.

AI manager

Via email, I asked one of the authors, Tomas Chamorro-Premuzic, a professor of business psychology, to give me a sense of how an AI boss might work.

He wrote: "It is not as far fetched as we may think. Computers and algorithms already have a lot of data on us. Imagine turning that data into accurate insights about us: who we are, what we want, how we are feeling, and what we need to do. Then imagine that system giving us instructions: to buy, to sell, to e-mail someone, to work on something. Likewise, such a system would be able to give us feedback on our performance. What I just described is an AI manager. If you want a robot manager or 'roboss,' all you need is a hardware or physical body around the AI."

He said such a system would work best when combined with human intelligence, so workers wouldn't report only to a computer.

The professor agreed an agile workforce would be well suited for this type of management: "First, most people switch from traditional employment to self-employment to avoid having a boss, yet that doesn't mean they are able to manage themselves. Second, gig-workers tend to be more technologically savvy, so you can see them being early adopters in the field. And third, AI bosses will always be much cheaper than having a human boss."

The future will be interesting, folks. We'll all be agile, and country music songs will include twangy gripes about how "my robot boss is a jerk".