Q&A: Hello Rabbi Michael
Hello Rabbi Michael
Question
How are you, sir?
I don’t know whether you remember our brief correspondence about the autonomous-vehicle research we are conducting.
Well, we’re making progress. If you’d like, it would be an honor for me to share with you materials that are already at advanced stages.
Allow me to ask your advice on a specific point.
One of the voices heard in the emerging discussion about development policy for autonomous vehicle systems in dealing with moral issues argues that the right thing in this context is to avoid instructing the vehicle’s operating system to act one way or another in such situations. According to this approach, the developers should feed the system a principled instruction telling it, “Never intentionally harm human beings,” and nothing more; alternatively, they should design a mechanism that operates almost like fate, in a random fashion, expressing the best equal chance for every possible party involved, in cases where that option is relevant.
How would you address this from a Torah perspective? Is it enough to refer to your article “Is Jewish Law Pluralistic?”
Sincerely,
Answer
Warm greetings.
I’m not sure I remember. Are you a doctoral student at the Hebrew University?
Reading a great deal of material is difficult for me because of the press of time, even though the topic interests me very much. If it’s possible to send material that can be read fairly quickly, or to describe the relevant points to me orally, I would greatly prefer that.
Clearly my article is not enough, and I’m not even sure it is relevant to this discussion (except perhaps in a very, very basic sense).
As for your question, I need a bit more detail about what exactly this instruction means and what it would do in specific situations, with a few representative examples. In general, when human beings are involved, the consideration of how many lives you save or lose is not always the relevant one. For example, in the trolley dilemma in Jewish law, according to most halakhic decisors there is no permission to divert the train toward one person in order to save many. On the other hand, when you are programming an automatic system, it is not a human being, and therefore it does make sense to program it so that as many lives as possible are saved, or as few people as possible are harmed. Even if there would be a “transgression” here, the offender is the vehicle (the programmer is only indirect causation). That itself is a fascinating question—whether there is a difference in the relevant considerations between programming a vehicle and actual behavior—and I would be glad to think about it a bit.
In general, the topic seems truly fascinating to me, and the question is how this can be done in a way that will not require time I do not have. Any suggestions? If you want to do real work, then perhaps we could think about something more organized and intensive if you have a budget to devote to it. If this is about general consultation and sporadic thinking, then we need to think of a more reasonable and focused format—focus on a specific question and try to think about it.
Discussion on Answer
Warm greetings.
There was some doctoral student (I think in communications) who corresponded with me on this topic, and that is why I asked. To my regret, I do not remember our correspondence.
The sentence you sent seems tautological to me. If Jewish law has a position, then by definition it does not allow different conduct. What needs justification here?
The question whether programming is subject to the rules of human behavior is another question, which I addressed briefly in the previous answer. Are you asking how to justify that? That is the interesting question here, and I have a few initial ideas. But you write that you already addressed it.
As for the book, the persecutors wrote a 700-page book in English (they still haven’t repented after seventy years. Hadn’t the Jewish people suffered enough? as our rabbis in HaHamishia HaKamerit said). That’s too much for me.
The Jews who worked on translating this book and wondering what exactly the knights of political correctness are actually saying in their convoluted formulations (“Whether and to what extent we should give weight, when it comes to preferring harm among animals, to the fact that there are animals more beloved to human beings and others less so, or whether human preferences are irrelevant when it comes to animals’ right to life”—the translator laughs and I pay) — apparently have not suffered enough…
And seriously: in principle you are surely right. And yet there is something sensible in the claim that it is preferable to evade the discussion rather than formulate principles for all of human society. Pretension is the name of the game.
Even if there is no real question here, I will still need to formulate a sentence that will not grate too much on the eyes of our target audience. But that is already my problem.
As for the discussion you mentioned: with God’s help, I will soon send you material (less than 700 pages), in the hope that you will find time to look it over, or that curiosity will win, or both.
And while we’re already talking, allow me to bless you with a hearty “more power to you” for the important article you were party to, about the ruling of Rabbi Yosef Messas of blessed memory regarding head-covering for women nowadays. Few articles make such a contribution in so many different respects.
In friendship,
Rabbi Michael,
First, I am merely a kollel fellow currently sheltered under the wings of Rabbi Asher Weiss and the Torah-and-technology research institute under his presidency (this is the second study he has produced).
Second, I have indeed conducted a comprehensive discussion weighing the considerations you raised (and quite a few others). Your sensors did indeed detect correctly. This is a fascinating and remarkably innovative discussion (and I hope I am doing it properly).
My question to you concerned the stage before the discussion itself, where I survey several positions in the scant research that currently exists. One of them (manufactured by Germany of Edom) sought to evade these determinations out of a pluralistic approach that anxiously views morality as coercion de facto. And no, they do not have a concrete proposal for what should be done instead. They formulate a few general sentences (“The vehicle should take the step that most promotes safety,” and the like), and leave the rest in more or less random hands. The book is attached.
I would like to address this approach in a very general way. You write that your article may be relevant only in a very basic sense; perhaps from that perspective a few basic words really are needed? How does one justify a sentence that says, “Insofar as Jewish law can provide guidance on how to act, then as a binding system it does not permit different conduct/planning”?
Thank you for your ever-quick willingness to help those who sit at the gates, and of course for the website.