Robots

One of the few things that Andrew Yang and I have in common is that we both have about the same chance of becoming the next President of the United States. Despite polling far behind the frontrunners, however, Yang strikes me as in many ways the most original of all the would-be candidates vying for the Democratic nomination and also, and by far, the most tech-savvy. And I owe to him my renewed interest specifically in robotics and in the potential impact machines provided with artificial intelligence might one day—and, according to Yang, one day very soon—have on our American landscape.

When most people—or at least most people my age—think about robots they generally think of unreal ones: Rosie the Robot Maid from The Jetsons, C-3PO and R2-D2 from Star Wars, Robo-Cop and Wall-E from the movies named for them, Optimus Prime and Megatron from the I’ve-lost-track-of-how-many Transformer movies, and lots of other random androids and tin-plated automatons dished up by Hollywood to the American public for their cinematic degustation. Mostly, though, these robots are just souped-up metal versions of regular people who, just like their flesh-and-blood prototypes, vary dramatically in terms of the strength of their moral fiber: some are good and some are evil; some are adorable, while others are malevolent and seriously creepy; some can only manage to do what human beings have pre-programmed them to be able to accomplish, while others are able to strike out on their own and become autonomous, or at least autonomous-ish, actors on the world stage. But the key criterion the robots mentioned above all share is their non-existence: all are made-up creations intended specifically to entertain as characters in movies or on television shows and none of them is real.

For most people, then, robotics is merely the branch of theoretical science that provides the ideational underpinning that makes R2-D2 real enough to be depicted in a movie that bills itself as futuristic, but not completely fantastic. And that was what I thought as well.

Enter Andrew Yang, who opened my eyes to details of which I had no idea at all.

Yang talks about the entry of robotics into the economic mainstream, not as a semi-plausible plot for some futuristic science fiction movie, but as a “fourth industrial revolution” already well underway. (The first, stretching out from the end of the eighteenth century through the beginning of the nineteenth, was about mechanization. The second, coming at the end of the nineteenth century, had to do with the introduction of electrical power. The third, during the second half of the twentieth century, had to do with the advent of computer technology. And, at least according to Andrew Yang’s understanding, the advent of robotics will bring in its automated wake change just as total and societally transformational as in their day were the introduction of computers or the invention of the mechanical engine.) Nor can the numbers he cites be easily dismissed: the nation appears in the last decade alone to have lost almost five million jobs to robotic automation. And the advent of self-driving trucks—in effect, car-robots—will, so Yang, cost the nation another 8.5 million jobs if the number of soon-to-be unemployed truck drivers is added to the number of soon-to-be-unemployed workers in the various service industries that cater to truckers while on they are on the road away from home.  And Yang predicts that the lost off 13.5 million jobs is only the beginning because, in the end, the advent of robotics will totally, permanently, and irreversibly change the American workplace. We either will or will not be ready. But what we will not be able to do will be to stem it all off with wishful thinking any more than people a quarter-century ago could have possibly halted the adoption of computer technology in American offices no matter how sincere their desire might well have been to protect workers with no computer skills from losing their jobs.

And so, with my interest already more than merely piqued, I found myself drawn powerfully to an extremely interesting responsum about Artificial-Intelligence-related issues adopted by the Committee on Jewish Law and Standards last June. (The CJLS is the highest legal authority within the Conservative Movement and the ultimate arbiter of halakhic legality and illicitness.) Written by Rabbi Daniel Nevins, currently the dean of the Rabbinical School at the Jewish Theological Seminary, the paper has the tantalizing title “Halakhic Responses to Artificial Intelligence and Autonomous Machines” and is an excellent example of just the kind of incisive, well-researched writing that characterizes the CJLS at its best. At almost fifty dense pages, it’s a big read. And a lot of it is couched in technical language that will be of interest mostly only to rabbis and scholars. But the larger picture is one of a thoughtful legist trying to respond to something entirely new in the world by drawing from the wellsprings of history and attempting to find contemporary relevance in lessons developed long ago by people who wouldn’t have been able even to dream of C-3PO or Rosie the Robot, let alone to imagine them actually existing. And yet the halakhah—the general term for Jewish law in all its complexity, inventiveness, and perplexitude—has been mined in the past to find responses to all sorts of new things, including steam engines, hearing aids, computers, and space travel. So why not robotics?

The questions Rabbi Nevins sets out for himself to answer boil down to three basic queries.

One has to do with the question of agency: can an intelligent machine able to make autonomous decisions be considered the author of its own deeds or must the responsibility of whatever R2-D2 does be laid at the feet of his original programmer?

A second has to do with ethics: should autonomous, thinking machines, including those programmed with the finest ethical principles, be permitted to make life-and-death decisions regarding human beings or should the ultimate responsibility for acting morally never be permitted to rest with machines—including those whose ability to weigh data and simultaneously to compare tens of thousands of precedents far outpaces the analogous ability even the brightest and most learned human beings could possibly cultivate?

And the third has to do with religion in general and with Judaism in specific, and asks whether a robot—or any autonomous, intelligent machine—can perform a mitzvah or utter a prayer either on somebody else’s behalf or, even more weirdly to consider, on its own behalf.

So those are Rabbi Nevins’s three core issues. Each in its own way is a refocus of the single basic question that underlies them all, however: can a machine capable of acting autonomously be taken seriously (or ethically or legally) as a person? To push that envelope just slightly further, I could ask if such a machine—or rather, once personhood is in some way deemed to inhere in the warp and woof of its existence, if such a “person”—could be deemed a Jew. Or, for that matter, if such a “person” could be supposed to possess any of the factors that we use to distinguish between different varieties of flesh-and-blood people like gender, nationality, ethnicity, race, sexual orientation, etc. Can a robot be a black person or a gay person? Can a robot be a man or a woman? This suddenly feels a lot more complicated than it seemed on the Jetsons!

Rabbi Nevins deals with all these issues intelligently and adroitly. (To read his responsum in full, click here.) And then, towards the end of his paper, he gets finally to the section that strikes me as being the crux of the matter, the one entitled “Androids as Religious Agents.”.

He begins by citing books by Gershom Scholem, Moshe Idel, Byron Sherwin about the concept of the golem, the man-made creature that entered halakhic discourse in the seventeenth century. And then he turns to the sources themselves.

The Sefer Yetzirah, generally considered the oldest extant book of Jewish mystical speculation, apparently already—and this is a very old book we’re talking about, one that some date as early as the second century CE—imagined the possibility that the scriptural reference to the souls that Abraham “made” in Haran was meant to be taken literally and that Abraham actually knew how to create what we would call an android—a kind of artificial human being lacking only speech and the kind of innate intelligence that can only come as a gift from God. And, indeed, that idea that in the righteous individual could conceivably inhere the ability artificially to create a living creature who would then lack only speech is already present in the Talmud, where we read that Rava, one of the masters of rabbinic Judaism in late antiquity, actually did create a man, albeit one who could not speak. And his remark that, if they were to wish it, “the righteous could create a whole world” of living creatures is also recorded, and in that same talmudic passage.

These passages were eventually taken seriously. The eminent halakhist, Rabbi Tzvi Ashkenazi (called the Ḥakham Tzvi, 1660-1718), for example, actually penned a scholarly responsum dealing with the question of whether the kind of person created artificially could be counted in a minyan, in a prayer quorum. (His answer was no.) His son, the even more famous Rabbi Jacob Emden (1697-1776), also took up the matter and determined that the speechless android is less like a mute human being than like an animal in human form—and so his answer was also no. Scholem and Idel discuss these sources and many others, but it was the late Rabbi Sherwin who apparently first realized that these texts could reasonably form the basis for a halakhic approach to technology in our day. Indeed, his 2004 book, Golems Among Us: How a Jewish Legend Can Help Us Navigate the Biotech Century, is still in print and is widely available. I recommend it to my readers highly.

Nevins spends time with all sorts of authors I haven’t read, people like Giulio Tononi and Michael Graziano who write about the complex interrelationship of consciousness, technology, and humanness—and thus about the nature of personhood itself, about what it means to be a person. He understands clearly that thinking about thinking machines is a way of thinking about what it means to be alive, what it means to be a human being, even what it means to exist at all. To imagine a world populated both by regular human beings and by the kind of androids depicted in the recent HBO hit series Westworld is simple enough. But to follow that thought through and attempt to imagine how civil rights and ethical prerogatives might inhere differently in born-people and made-people is, to say the least, daunting. 

Andrew Yang is personally responsible for bringing this issue to the national stage and we should thank him for that. Daniel Nevins has effectively shown that there is more than enough water in ancient wellsprings from which scholars can and should drink as they ponder these abstruse, confusing issues, that he too deserves our thanks. But where exactly this will all take us—that, at least as far as I can see—is still entirely up in the air.