Autonomous Driving — Where are the Psychologists?
“ The fig tree is pollinated only by the insect Blasto-phaga grossorum. The larva of the insect lives in the ovary of the fig tree, and there it gets its food. The tree and the insect, are thus heavily interdependent: the tree cannot reproduce without the insect; the insect cannot eat without the tree; together, they constitute not only a viable but a productive and thriving partnership. This cooperative “living together in intimate association, or even close union, of two dissimilar organisms” is called symbiosis…The purposes of this paper are to present the concept of and hopefully, to foster the development of man-computer symbiosis”
In his seminal paper, Man-Computer Symbiosis, J.C.R Licklider (Lick) laid out his vision of the relationship between man and computer. Up to that point, many viewed the computer as simply a tool to execute algorithms. What Lick envisioned was quite different. Rather than a tool to be used, Lick imagined a relationship to be had — a friend that could help solve problems that otherwise not be solved alone. “Humans would set the goals, supply the motivations, develop the hypotheses, and ask the questions. Computers would simulate the mechanisms and models, carry out the procedures, interpolate, extrapolate and transform.” His vision and subsequent leadership position in ARPA, paved the way for the next several decades of research that eventually led to the internet and personal computing revolution.
Today we find ourselves at a similar junction — one in which the relationship between man and computer must be reconsidered. This time man is not ‘partnering’ with machine to enhance his own capacity. Rather, we are developing the computer that needs to understand man to inform and guide the development of new AI products.
Autonomous driving is the best example. Several prominent AV executives (e.g., Amnon Shashua, Gill Pratt) have gone on record saying that the hardest part of ‘cracking the autonomous driving code’ is driving policy — figuring out how to teach a machine to negotiate in a multi-player game. Sensing and mapping aren’t enough to maintain safety when the machine has to anticipate the actions of irrational and unpredictable players (humans). Therefore without a deep understanding of humans, machines will never be able to realize the gains that many suggest the might.
Given that understanding human psychology is a fundamental tenet to ‘cracking the code’, why aren’t AV startups hiring psychologists? As I scan through the hiring packages of Drive.ai, Nutonomy, Aurora, Argo and others, I see hundreds of postings for machine experts — engineers — but not a single attempt to source human experts — psychologists. These companies are solving single-agent problems in a multi-agent world — and I portend that until they realize that err, the market penetration and AV adoption curves will significantly under-perform forecasts. It took the genius (and psychologist) J.C.R Licklider to foresee and pioneer interactive computing and will likely require a similar background to really foresee and pioneer autonomous driving.