Words&Dirt

John Markoff’s Machines of Loving Grace: The Quest for Common Ground Between Humans and Robots is another addition to the growing stack of books designed to help us think about the relationship between humanity and emerging technologies. Markoff offers a detailed history of artificial intelligence and robotics, and attempts to show how past trends are relevant (or not) to the modern moment. Although the book touches on some important themes, it’s significantly inferior to other texts I’ve read in this field (e.g. recent works by Martin Ford, Nick Bostrom, James Barrat, and Ted Chu).

The focal point of Markoff’s project is the paradox of automation: “The same technologies that extend the intellectual power of humans can displace them as well” (xii). He draws a sharp distinction between artificial intelligence (AI) and intelligence augmentation (IA), claiming that advances in AI tend to replace human labor, whereas IA usually amplifies the abilities of human workers. The usefulness of this dichotomy is dubious because there’s no reason why AI must necessarily displace human labor while IA doesn’t (e.g. a breakthrough in AI could create an entirely new job market for humans, and a breakthrough in IA could make it possible for one human to do the work of ten).  Markoff comes down firmly on the side of IA, and he’s not alone: many top minds in AI research have forsaken it for the IA camp in recent years.

Complementing Markoff’s thesis are a host of worthy topics, including the backgrounds and values of technology designers, growing worries about technological unemployment, the financial relationship between technology companies and the US military, and reformulations of human identity that will take place in an age of increasingly autonomous machines. In general, Markoff looks back more than he looks ahead; Machines of Loving Grace is heavy on facts and proper nouns, but light on ideas and insights that might prove useful moving forward.

The strongest and freshest argument here is that building a human element into new technologies is primarily a design choice, and not just a matter of efficiency or convenience. The views of designers, then, matter a great deal:

Despite the growing debate over the consequences of the next generation of automation, there has been very little discussion about the designers and their values. When pressed, the computer scientists, roboticists, and technologists offer conflicting views. Some want to replace humans with machines; some are resigned to the inevitability…and some of them just as passionately want to build machines to extend the reach of humans. (26) 

Since discussions of technological “progress” have a tendency to reach absurd levels of abstraction, Markoff’s determination to put a human face on our technological vanguard is welcome. He outlines the early influences and careers of a long list of technologists, demonstrating an impressive breadth of research. The reader gains a concrete sense of the human stories and motivations behind some of our most ubiquitous tools.

This approach, however, has some significant downsides. Markoff seems to have fallen down the Silicon Valley rabbit hole, at least partially buying into the cults of personality for which it has become so notorious. In addition to highlighting the values and formative experiences of technologists, Markoff also wastes many pages describing obscure projects that went nowhere and inane details about who worked for what company or knew the right people in the right moment. Here’s a typical example:

In 1999, [Andy] Rubin started Palo Alto-based Danger, Inc., a smartphone handset maker, with two close friends who had also been Apple engineers. The company name reflected Rubin’s early obsession with robots. (In the 1960s science fiction television series Lost in Space, a robot guardian for a young boy would say “Danger, Will Robinson!” whenever trouble loomed.) Danger created an early smartphone called the Sidekick, which was released in 2002. It attracted a diverse cult following with its switchblade-style slide-out keyboard, downloadable software, email, and backups of personal information in “the cloud.” While most businesspeople were still chained to their BlackBerrys, the Sidekick found popularity among young people and hipsters, many of whom switched from PalmPilots. (240-1)

Ugh. Who gives a shit? There’s a reason the Sidekick isn’t a household name. This “inside” information is irrelevant for anyone who is not a technology entrepreneur or historian, and contributes nothing to Markoff’s ostensible goals. This summary of Rubin’s career could easily have been condensed down to a sentence or two, or excised entirely. The odd paragraph like this wouldn’t be so bad, but the majority of the book is padded with details that trump perspective. Markoff’s “one damn thing after another” style is stifling––even the sections on topics that interested me were boring to read.

If surfeits of superfluous detail and lifeless prose were my biggest gripes, I’d give this book a higher rating. But there are much more troubling issues, the first of which goes back to Markoff’s fuzzy distinction between AI and IA. These labels are useful to an extent, and there are definitely important differences between the goals and practices of AI and IA researchers. Markoff points out that the two fields don’t overlap as much as one might expect (282). Even so, I’m not convinced the dichotomy cashes out the way Markoff wants it to. His final characterization of the situation is telling:

Whether computing technologies are deployed to extend human capabilities or to replace them is more a consequence of the particular economic system in which they are created and used than anything inherent in the technologies themselves. In a capitalist economy, if artificial intelligence technologies improve to the point that they can replace new kinds of white-collar and professional workers, they will inevitably be used in that way. That lesson carries forward in the differing approaches of the software engineers, AI researchers, roboticists, and hackers who are the designers of these future systems. It should be obvious that Bill Joy’s warning that ‘the future doesn’t need us’ is just one possible outcome. It is equally apparent that the world transformed by these technologies doesn’t have to play out catastrophically. (342-3)

This is an adequate summary of the current technological moment, and I appreciate Markoff’s attempt to point out the possible ways the AI/IA revolution might pan out. But this perspective also contradicts the dichotomy put forth in the rest of the text. Markoff consistently draws a bright line between a future where machines replace human labor entirely and one where they make us better at our jobs. But it should be clear to anyone paying attention that both of these scenarios are inevitable. AI and IA will both have critical roles in the economies of the future. Furthermore, a future that “doesn’t need us” isn’t necessarily a catastrophe. Humanity might give up its status as Earth’s apex intelligence, but that doesn’t mean doom in all cases. Beyond many possible positive outcomes for humans, this might also mean a reprieve for our long-abused natural environments. A future that runs more like a self-driving car than a human-operated one could be better in countless ways.

By far the weakest aspect of Machines of Loving Grace is its sanguine attitude about technological unemployment. Markoff seems content to think that the IA camp will save us by extending the capabilities of human workers, but fails to realize that that process is a hallmark of the “do more with less” outsourcing culture non-elite workers have been battling for decades. The prevailing order demonstrates that when one human can do more, fewer humans are necessary. CEOs and shareholders cheer, jobs evaporate, and remaining workers have to work harder, which they do without complaint for fear that they too will become dispensable. So regardless of whether we decide to favor AI or IA, technological unemployment will still be a major problem. (For an up-to-date look at wages and employment statistics, I recommend this recent podcast. And here’s an interview with someone who takes technological unemployment seriously and wants to find workable solutions.)

I suspect Markoff’s blind spot has to do with his reluctance to address the modern concept of “work.” He brushes off the arguments of Martin Ford and Jeremy Rifkin, who warn that we need to shake up our ideas about labor and what it means to “make a living” if we want a future that’s better for everyone, not just for the elites with whom Markoff is so smitten. It doesn’t help that almost all his interview subjects are wealthy workaholics who have nothing in common with people who have to work menial, mindless jobs just to get by.

Markoff is so caught up in the history of technology that he doesn’t seem to realize that it’s not very useful for forecasting the future––especially not in a time when technological progress is advancing at an unprecedented pace. Think of it this way: how useful would a book about the history of computers that was published before the advent of the Internet be for people in the post-Internet world? It probably wouldn’t amount to more than a historical curiosity, and its predictive power would be extremely limited, if not entirely worthless. We don’t know exactly where the next automation breakthrough(s) will come from, but it’s a good bet that the world will look very different shortly thereafter. If the perspectives put forth in Machines of Loving Grace are already questionable, how useful will they be then?

The sad irony is that we should be excited about technological unemployment. Why not focus on how AI and IA can positively affect human experience and communities, rather than obsess about how they can supercharge our overly-consumptive capitalist economies? We should welcome these developments as a chance to allow people to work less, have more time for their families and hobbies, and use technology to discover new realms of entertainment and pleasure. But to do this, we’ll have to decouple healthcare from employment, and also find ways to subsidize the living costs of people who can only find part-time work or whose jobs are gone for good. Markoff has nothing to say about these pressing matters, nor does he mention the role of government in this process (positive or otherwise).

All this indicates that Machines of Loving Grace doesn’t offer any practical answers to the problems it purports to address. It’s more of a historical overview peppered with Silicon Valley puff pieces. “This is about us,” Markoff concludes, “about humans and the kind of world we will create. It’s not about the machines” (344). Another trite dichotomy. The truth is that the future will be a joint product of the commerce and competition between humanity and our machines. It’s not only about us, nor is it only about the machines. The reality is far more complex.

Rating: 4/10