OpenAI Scholars 2019: Meet our Scholars

OpenAI has named its 2019 class of Scholars, whittling down a pool of 550 applicants to just eight. The program, which pays participants a stipend to study deep learning full-time with mentorship from OpenAI staff, is designed to pull talent from outside the usual computer-science pipeline into AI research.

What stands out about this cohort isn't the acceptance rate, brutal as it is at roughly 1.5%. It's the résumés. OpenAI says the group's collective expertise spans literature, philosophy, cell biology, statistics, economics, quantum physics, and business innovation. That's not a lineup you'd expect from a typical machine learning bootcamp, and it's clearly the point.

OpenAI has run versions of this program before, mostly aimed at increasing diversity in a field that skews heavily toward a narrow set of academic backgrounds. The bet here seems straightforward: people who've spent years thinking rigorously about ethics, markets, or cellular systems might bring pattern-recognition instincts that a standard CS degree doesn't teach. Whether that translates into better research output is the kind of thing that's hard to measure directly, but the logic tracks with a broader industry realization that AI's hardest problems, alignment, bias, interpretability, aren't purely technical.

The scholars will spend months building projects and presumably publishing writeups, which has been the pattern in past cohorts. For OpenAI, it's a relatively cheap way to test whether nontraditional backgrounds produce nontraditional insights. For the eight chosen, it's a shot at breaking into one of the most competitive corners of the tech industry without the usual credentialing.