Employment is in chaos. Tariffs are battering global supply chains. The tech companies behind large language models (LLMs) say their products will soon automate away huge swaths of jobs, replacing occupations that have existed for decades (like programming) or millennia (writing, child care) with unprecedented new jobs overseeing automated work — jobs whose skill needs no one can predict. If the generative AI geese do lay these golden eggs for investors, we may face unemployment on a scale to rival the industrial revolution. If they don’t, some already fear a tech-stock crash to rival 1929. Either way, the promises of LLMs are already causing mass firings and turning employment on its head. Last year, graduates with bachelor degrees in art history and philosophy secured jobs at better rates than those with computer-science degrees, the proverbial most and least employable majors trading places.
This is a dire moment to enter the work force.
It is a perfect moment to spend four to seven years acquiring rare and valuable skills before entering a work force which will not be, in five years, what it is now.
As an alumna of Harvard’s Ph.D. program, it is heartbreaking watching my alma mater, which just triumphantly won its standoff with President Trump, cut its Ph.D. admissions by 75 percent when doctorates have never been more needed.
Proponents of slashing Ph.D. programs tend to focus on the claim that there are no academic jobs, therefore no use for doctorates. Now, academic hiring is still happening; indeed, my own department just conducted a deep survey of our Ph.D. alums which confirmed that the percentage of them securing academic jobs has remained shockingly constant at about 70 percent all the way back to 1981. But rather than debating the numbers, for the moment let’s explore the doomsday scenario in which the worst fears are realized and there is zero academic-job hiring in 2030.
The Ph.D. process teaches broad expertise in research methods. Doctorate holders know how to evaluate evidence, chase down sources, verify claims. They learn methods for finding reliable information, when to be suspicious of data, and how to act upon those suspicions. In addition to learning what we know, a Ph.D. student gains expertise in what we don’t know, learning where the frontiers in a field are, the big questions, incompleteness, and areas a field has not yet tackled. Someone with a Ph.D. can often look at a graph and say, not “that isn’t true,” but “that data doesn’t exist.”
This is a dire moment to enter the work force. It is a perfect moment to spend four to seven years acquiring rare and valuable skills before entering a work force which will not be, in five years, what it is now.
For the first half of the 2010s, Nokia hired people with graduate degrees in any subject, from art history to linguistics to astrogeology, to verify map coordinates for its navigation program Nokia HERE, because supervisors found that people with a master’s degree or doctorate in any subject had the right combination of rigor, patience, and research experience to do the thorough work the project needed. That kind of work — checking the output of an automated computer-mapping process — is extremely similar to what might shortly become the most in-demand new occupation in a decade. We should not greet an age likely to face an unprecedented need for fact-checking and verification by ceasing to train the very experts who can best meet that demand.
One crisis facing us is the mass erosion of confidence in expertise. Daily we see political leaders purge experts from public service, replace skilled specialists with businessmen, and even paint expertise in a field as a disqualifying bias. Anti-science rhetoric has stirred even the most neutral and apolitical scientific bodies to sound the alarm. Ph.D. holders are not just experts but ambassadors for expertise, able to explain and show what real expertise means in a field, and help people see why expertise is valuable. Throughout the last century, Ph.D. holders have worked in government, journalism, law, social service, industry, justice, education, defense, arts, entertainment; we cheer to see a Ph.D. holder running for office, advising Congress, staffing the halls of power, or speaking on TV.
Every economic field has found the rare Ph.D. holder who entered it to bring invaluable research skills and perspective. In my own world of science-fiction and fantasy novel-writing, a huge range of Ph.D. holders like myself — Arkady Martine (history), Amitav Ghosh (social anthropology), R.F. Kuang (East Asian languages and literature), Alastair Reynolds (astrophysics), Harry Turtledove (history), Alison Sinclair (epidemiology) — are pouring out award-winning books saturated with our individual expertise, and bringing enthusiasm for our fields to tens of thousands of readers. Amid an active effort to purge expertise from the halls of power, it is baffling to see colleges themselves climb on that bandwagon and accept the claim that, if there are no academic jobs, we should train no Ph.D.s — a claim which accepts the false premise that there is no use for Ph.D.s outside academe.
Earth’s oldest universities will soon celebrate their 1,000-year anniversary. Our collective project of advancing and safeguarding the future of human industry has faced seismic historical shifts before, from the print revolution on. We have never, in that history, doubted that producing expertise is valuable for the world, not only in tried-and-true ways but in new ways which would help meet and lead the changes creating new eras. If we fear computer-generated falsehoods may soon drown out both truth and real human expression, how will we develop new means to meet this crisis if we respond to it by shrinking Ph.D. cohorts and producing fewer experts capable of helping us face this revolution?
We do not know what jobs will be in demand in five years, or in one year, and undergraduates now choosing majors based on employability face the terrifying likelihood that what is employable today will not be so by the time they graduate. Doctoral study is the inverse of this uncertainty. We absolutely know that expertise, research skills, patience, and understanding how to go about investigating and evaluating truth and falsehood will always be in demand, especially in a society racing to meet new needs in an information revolution. There has never been a better time, from the student’s perspective, to continue studying for a few more years to gain deep skills. Nor has there ever been a better time for institutions to support the production of deeply expert leaders who will be needed in the next stage of this crisis.
We did not respond to past revolutions — political or technological — by doubting our core mission to safeguard and advance the future of human inquiry. We should not do so now.