Intelligence Work
What we associate with ‘intelligence’ in machines is the work of computation; work that was first done by humans. Lorraine Daston assembles a history of calculation since the mid 1700s in Europe (pdf), detailing how the “hard labour” of mathematical calculation in physics, nautical navigation and astronomy was performed. It was achieved by a pyramid-like hierarchy as follows: a few mathematicians on top, who determined the formulae and designed the logarithms; below them, the algebraists, who translated the logarithms into numerical form; and below them, roughly seventy or eighty “workers”, who only knew elementary arithmetic but “actually performed the millions of additions and subtractions and entered them by hand into seventeen folio volumes.” Daston notes that it was the humans toiling away mechanically that sparked the idea for Charles Babbage’s Difference Engine, the precursor to the modern computer.
Speaking of the masses of workers at the bottom of the calculation pyramid, Daston writes that Babbage was known to often say that “the fewest errors came particularly from those who had the most limited intelligence, [who had] an automatic existence, so to speak.” (17) In other words, it was possible to imagine humans as being ‘automatic’. Fast forward a few decades and the mechanical, time-consuming labour of calculation was outsourced to women, who were considerably cheaper to hire as workers and treat as machines.
Jennifer Light’s history of computers shows that till the middle of the 20th century, ‘computer’ was, quite literally, how women calculators were referred to in their place of work; it was their designation, like ‘typist’. The complex and tedious work of calculation required of these ‘computers’ was rarely, if ever, acknowledged. It is only very recently that the work of Katherine Johnson and her team of mathematicians, whose work was integral to the success of NASA’s first human spaceflight project, came to light in the book (and subsequent film) Hidden Figures. Over time, women’s work has been systematically erased from the history of computers, and women pushed out of the technology industry. Marie Hicks’ book Programmed Inequality tells a similar story: of how Britain lost its edge in computing because it excluded women engineers and mathematicians.
Increased automation generates anxieties associated with the changing role of the human. However, the mechanisation of calculation with machines in the 19th and 20th centuries did not necessarily erase the involvement of humans. If anything, a more daunting task arose: coordinating and apportioning work between humans and machines; getting them to ‘flow’ together; figuring out who (or what) was better at which kind of task. So as calculation and calculating devices permeated industries like insurance, government departments, colonial management companies and military and operations research, the algebraists became middle managers who oversaw tasks between humans and machines.
Questions of human-machine interaction remain fresh today as we negotiate, for example, handover between human drivers and auto-pilot systems in driverless cars, for example. Or have humans tag images, like in CAPTCHAs, so that machine learning systems learn to identify objects better. The work of intelligence thus continues as an interaction between human and machine.