An IBM executive told a recent conference that when supercomputer Deep Blue was halfway through its 1997 chess match with Garry Kasparov, it made a random move, due to a software bug. Assuming the machine was smarter than it was, Kasparov later made a strategic error that helped hand Deep Blue victory in the match.
Thomas Davenport and Julia Kirby warn that humans could, like Kasparov, cede the future to machines too easily. “Many knowledge workers are fearful,” they write. “We should be concerned, given the potential for these unprecedented tools to make us redundant. But we should not feel helpless in the midst of the large-scale change unfolding around us.”
Only Humans Need Apply falls into the sub-genre of techno-optimism, at the opposite end of the bookshelf from, say, Martin Ford’s doomier Rise of the Robots .
Not all of every job can be automated, the authors point out. Change may be a long time coming (they cite futurist Paul Saffo’s wise insight: “Never mistake a clear view for a short distance”). Critically, humans have the power to shape their destiny. They can develop, program and direct machines, and adapt to working with them, in a process of “augmentation” rather than mere automation. Machines, up to now “the brawn to our brains, can become the brains to our brio”.
Having deftly outlined the gloomier forecasts, Kirby and Davenport lay out how to avoid submission to, or substitution by, machines. A few senior managers will step up into overseeing automated systems. Other workers will step aside, by developing careers in areas machines are not as good at, such as motivation, creativity or empathy.
Still others will step in (by learning more about how computers work and how to improve them), step narrowly (becoming super-specialists), or step forward to develop new systems and technology.
Plenty of people have already made these changes, and they describe many of them. Here is the redundant lawyer who “stepped in” to become an expert overseer of automated contract reviews; or the editor who “stepped up” to develop a computerised system for preparing and publishing sports statistics and earnings reports. If you pair adaptable humans with computers, you can remove tedious tasks and manage the change humanely.
Developers must “look for ways to help humans perform their most human and most valuable work better”, the authors say. Simply asserting this bright future will not make it occur, however. The temptation for bosses to opt for the most efficient robot solution is strong. Davenport and Kirby “expect corporations’ efforts to keep human workers employable will become part of their ‘social licence to operate’”. But is that wishful thinking and, even if not, will it happen soon enough?
This is a fine call to action in the face of uncertainty. But let us program the automated story-writing tool for a sequel in 2026. Only then will we know whether humans who stepped up, in or forward, sustained a fulfilling working existence, or were merely treading water ahead of a tsunami.