HR tech: when the computer says no

Employees have long grumbled about receiving unfair treatment from their bosses. They may soon have to get used to receiving it from a computer instead. The use of artificial intelligence to make hiring and firing decisions is already with us, in some sectors. Trade unions are concerned. So should companies that provide HR services, like recruiters Adecco, Manpower and Randstad.

In the past, AI’s role in HR was confined to boring, repetitive tasks, using tools that were something like databases on steroids. New-generation AI will expand the activities machines can undertake.

The snag is that AI is only as good as the data it feeds on. And HR data tends to be patchy, heavily context-dependent and riddled with bias. Flaws in the analysis are hard to spot, too. Self-taught machines are apparently not keen to show their workings. Hence the TUC’s concern and the efforts to make AI more transparent.

While the technology is hardly perfect, investors can see its potential. Private investment in HR Tech companies more than doubled in 2022, to $1.6bn, according to Stanford’s latest Artificial Intelligence Index Report. That’s a big jump, but still low compared to sectors in which AI got an earlier start, such as medicine, which brought in $6bn.

There is some room for growth. AI has been adopted in HR functions in just over a tenth of industries McKinsey reviewed. That is about half the penetration in business units such as Risk or Corporate Strategy.

With disruptive HR tech knocking at the door, players are jostling for position. SAP, Oracle and Workday, which already provide managerial technology and collect company data, should find new business opportunities. One-trick ponies, such as recruitment specialists HireVue, or Beamary — may end up getting snapped up.

Traditional recruiters such as Manpower may have a challenge. Their business model offers low-cost manpower for their clients. Initially, AI may be a tailwind, reducing costs further. But as companies adopt their own systems, any advantage will be eroded.

In-house HR people, too, will need to find a new roles, perhaps to audit the results of the AI tool, and provide an appeals system for those who feel wronged by any algorithms. Uncomfortable though we may feel with human errors of judgment, they may still be preferable to those by a machine.