We need to talk about automation.

Unless you used to work in a steel mill, the prospect of losing your job to a robot might sound like the dissociated musing of some science fiction writer. For most Americans, automation is like the nuclear fusion of the job market: it’s always coming, but it’s never quite here.

That might already be changing.

Last week gave the world its first case of an automated car killing a pedestrian, when a self-driving Uber was unable to avoid hitting a woman crossing the street in Tempe, Arizona on Sunday, March 18. Subsequent reports have suggested the crash was unavoidable, and Uber has responded by halting, but not necessarily discontinuing, its self-driving car tests in major metropolitan areas.

The key takeaway from this tragedy is in the public and private sector’s reaction. The world’s first death by automated vehicle is largely being treated as a bump in the road to self-driving cars and the benefits they promise. If death can’t stop automation, nothing can. 

Automation’s effects are predicted to expand dramatically in the coming years. A 2017 report from the McKinsey Global Institute think tank anticipated that by 2030, up to 800 million people — one-fifth of the world’s workforce — will lose their jobs to a machine.

Researchers from Oxford University released a report in 2013 that ranked over 700 occupations in order of their ability to be automated. Twelve of those professions, including cargo agents, sewing and telemarketing, were given a 99 percent chance of computerization, while over 400 jobs found themselves over the 50 percent mark. Predictably, the jobs rated as least doable by machines, like healthcare social work (.035 percent) or general dentistry (.044 percent) involve a great deal of thought, empathy or physical precision, qualities computers only possess in sci-fi movies.

But is it really safe to assume that machines won’t be able to make up the gap in these areas soon?

If a computer’s capability is only limited by its programming and processing power, then surely automation’s reach will extend far beyond the assembly line before long. There must be an algorithm complex enough to settle an HR dispute, to write a heart-wrenching novel or a nuanced opinion article. If we’re only two years away from roads full of self-driving cars, like Forbes predicted, then who can anticipate what 30 or 40 years of development will bring?

After all, it only took one lifetime for computers to go from room-sized adding machines to the phones and laptops you can’t imagine your life without.

So where does all that leave the young workers and college students who can expect to live out their whole working lives beneath technology’s growing shadow? If computers can do my job or your job badly now, will they do it better than us before we can retire? Will the skill sets we acquire in college even be relevant in a decade?

If we want answers to questions about how computerization will change our world, it might be best to ask the people who have grappled with these questions the longest: computer scientists.

I.V. Ramakrishnan, a Stony Brook professor and associate dean of the department of computer science, has been adapting to increasingly complicated computers since he entered the industry nearly four decades ago. His advice to educators struggling to teach in a changing world is simple: teach the core principles, not the latest techniques.

“In my department we don’t teach [students] the latest technology,” Ramakrishnan said. “What we teach them are the principles. Principles are always long lasting. Technology will come today and it won’t be there tomorrow… As long as there is love for learning, they will be ok.”

If the history of technological development has proven anything, it’s that new technologies ultimately create more jobs than they alter or destroy. In a sense, automation is a centuries old issue dating back to the days of the first factories. Before the textile industry became mechanized, it might have taken one person days to weave a basket or sew a shirt. The power loom might have taken seamstress’ jobs, but they created scores of factory jobs and even managed to introduce new skilled labor building and repairing machines.

Hindsight might one day prove that worrying about automation-induced unemployment was as short-sighted as worrying that newspapers would destroy the American social landscape. Sure, new technology has a way of replacing whatever came before it (I forget that I still have a landline) but the one constant is humanity’s ability to adapt to change.

We are all perfectly capable of learning to deal with changing circumstances, whether it’s a self-driving car or the automated factory that built it. If we weren’t, we would be dinosaurs.