In recent years, the job market has transformed, driven largely by advancements in technology. Both employers and job seekers are leveraging artificial intelligence (AI) to enhance their chances of success. Hiring companies utilize AI to meticulously analyze numerous resumes, searching for the ideal candidate. Meanwhile, applicants are harnessing similar tools to tailor their resumes to fit specific job descriptions. Despite the common objective, this dynamic isn’t as new as it appears. Greg Downey, the Evjue-Bascom Professor of Journalism and Mass Communication and Director of the Information School, reveals insights into this ongoing evolution.
“We’ve been here before, even if we didn’t call it AI, or even if it wasn’t computers,” remarks Downey. “There’s a history of how people think about connecting individuals with careers, utilizing tools, algorithms, and technologies that have existed for the past century.”
In addition to his teaching roles, Downey spent ten years offering a career course to undergraduates through SuccessWorks, the career advising service of the College of Letters & Science. His experiences there ignited his interest in exploring the historical roots of technological approaches to career selection. He discussed his findings at the 2026 CultureCon, an AI summit held in Madison from April 21-23.
Downey’s research traces back to the 1920s, specifically to Clark Hull, a psychology professor at UW–Madison until 1929. Hull was a pioneer in aptitude testing, envisioning a set of approximately 20 tests that could identify which students would thrive in various professions, ranging from telegraph operators to streetcar drivers. He even created a basic machine to process the test results.
“There were no computers, there was no AI,” states Downey. “My assertion is we were attempting to create a system or machine to predict job success long before the technology existed.”
Although Hull’s concept was innovative, it proved impractical. The main challenge wasn’t the design of the ideal tests; it was the time and expense involved in administering and interpreting them. Each occupation required a unique formula, complicating matters further.
Hull’s work marked only the beginning. Fast forward to the 1960s, where a group of psychologists in Pittsburgh launched Project Talent, influencing around 4,000 schools across the United States to adopt a two-day testing initiative aimed at matching student aptitudes with potential career opportunities. By this time, computers were able to handle the vast amounts of data being collected, allowing results to be packaged as career guidance tools for school districts. Even decades later, variations of this test remained in use, despite their reliance on increasingly outdated data.
“Those involved at any point in this history did not have bad intentions,” Downey explains. “They operated under the biases of their era, driven by a desire to improve outcomes. They sought technological solutions, a quest that continues to be enticing.”
However, modern technology also has its limitations. Just over a decade ago, Amazon created an AI model aimed at identifying suitable candidates for its engineering and data science teams. This model was trained on the resumes and job histories of successful employees but yielded significantly biased results.
“Their AI predicted it should only hire men,” Downey points out. “They were drawing from a pool of predominantly male candidates from specific tech programs and schools, hence the flawed perspective on what success genuinely looks like.”
Downey expresses concern that the rapid promotion and implementation of contemporary AI tools may not necessarily deliver cost savings and improved outcomes. Instead, he fears it could lead to an endless cycle of chasing the next technological breakthrough. He likens this situation to the 60s and 70s when experts predicted that computers would radically transform American education—a transformation that, while significant, fell short of the grand expectations.
“The historical lesson is that when it comes to technology, we often set our sights high and embark on grand initiatives, but very few achieve lasting success,” says Downey. “However, from each of these ambitious efforts, we glean insights that gradually get integrated into our practices. That’s what we should focus on.”
In conclusion, as we navigate the evolving landscape of job searching and hiring, it is essential to reflect on the lessons of the past. While technology offers unparalleled tools and possibilities, understanding its limitations and historical context will enable us to harness its potential effectively and ethically.
