In recent months, anthropologists have been rummaging through the grassroots of America’s workforce jungle in search of an answer to one of the great questions of our time: what happens to human jobs when robots arrive?

You might expect the answer to be very depressing. If there is one thing on which almost all economists agree, it is that digital technologies are performing many jobs once done by humans.

Manufacturing offers a particularly stark example of this. A study by Ball State university suggests that 5.6m US manufacturing jobs were lost between 2000 and 2010 — almost nine in 10 thanks to automation, not trade. It could be worse: McKinsey, a consultancy, estimates that 45 per cent of the tasks currently done by humans could be automated as the pattern spreads into the service sector. This equates to $2tn in annual wages — and millions of jobs.

That sounds scary. There is, however, an intriguing twist. When anthropologists have conducted “participation observation” among American workers — that is, observing at what is actually happening in people’s everyday lives, rather than looking at top-down statistics — they discovered a more complex story than the raw numbers suggest.

Yes, machines are wiping out some human jobs but people are also working with robots in new roles. That more upbeat story tends to be obscured, yet it deserves a great deal more attention — particularly when president-elect Donald Trump takes office next month.

Consider the findings of Benjamin Shestakofsky, an anthropologist who spent 19 months inside a California company that uses digital technologies to connect buyers and sellers of domestic services. Mr Shestakovsky initially assumed that his research would show how machines were replacing human workers. When he did grassroots analysis he realised that the company was growing so fast, with such big and complex computing systems, that it was constantly drafting more humans — not robots — to monitor, manage and interpret the data. “Software automation can substitute for labour but it also creates new human-machine complementaries,” he told an American Anthropological Association meeting recently, noting that companies “are creating new types of jobs”.

Shreeharsh Kelkar, another anthropologist, saw the same pattern in the education world. Until recently it was presumed that the rise of digital teaching tools would make human teachers less important. But watching educators in action, Mr Kelkar found that human teachers are working with these digital tools to be more efficient. The issue is not that computers are automating jobs away, he says, but that “assemblages of humans and computers are emerging”.

An obvious response is that it is far from clear whether these anecdotes are typical, nor does anyone know whether these new “assemblages” of human and machine will create enough jobs to offset those lost to automation. In addition, new digitised jobs may seem less attractive than the old roles since they are often structured as “contingent work”, with self-employed workers who provide services on demand.

Still, the findings of the anthropologists should not be ignored. For one thing, they suggest that there is a burning need for policymakers to obtain much better data on what is really happening in the American workplace. Anthropological studies are small scale, while the macro-level data are surprisingly weak, partly because the Bureau of Labor Statisticstends to collect data through traditional channels. “We don’t know what is going on with contingent work today,” says Mary Gray, an anthropologist who works at Microsoft. “The tech companies don’t track labour any better than the BLS.”

Second, if anyone does manage to paint an accurate portrait of that labour force, they need to show this to Mr Trump. In recent months the president-elect has repeatedly stated that he is determined to keep more manufacturing business in America, partly because he — wrongly — likes to blame the loss of manufacturing jobs to competition from China or Mexico. But if he does succeed in this goal of America First he will — paradoxically — only accelerate the automation trend as companies will scramble to cut costs. This is not necessarily a bad thing but it suggests that Mr Trump’s hopes of recreating old-style American jobs is wrong-headed.

That leads to the third point: the urgent need for a bigger policy debate about how to prepare workers for this new world. Workforce training needs to change to instil more digital skills. New types of social security, health and pension systems are necessary to accommodate contingent workers.

Some policymakers understand this. Senators such as Mark Warner, a Democrat, for example, are pushing for new safety nets for contingent workers. But if this debate is to secure any serious traction, it is imperative that the technology sector itself steps in. Hitherto, Silicon Valley has not been particularly vocal on these questions, but Mr Trump seems intent on pulling them into the spotlight: last week he summoned tech leaders to Trump Tower to “reassure” them about his plans.

So Silicon Valley should seize this chance and start a dialogue about how to help humans deal with all those robots in the workforce. Otherwise, the day will come when Silicon Valley itself could find itself being blamed for American job losses.