What happens to the People of the Book in an era when the book is replaced by the computer?
You may be familiar with the idea of Rabbi Google, or asking the Internet our Jewish questions and receiving an approximation of a response. But what happens when that response becomes much more robust, when the framework for asking the questions is no longer just a website that can send you to other sites but an intelligence that comes close to – or even exceeds – human abilities?
Artificial intelligence (AI) is poised to be the next big thing in information technology. Its proponents claim that AI will be able to learn like humans and will provide a human/computer interface that is more natural and conversational.
Many technologists even posit the possibility that an artificial intelligence will attain genuine consciousness and will need to be treated as a being. How will this new future affect how we live our Jewish lives?
I am not generally a futurist; I don’t like to make predictions about what will happen and what our world will look like. Nevertheless, it can be useful to think about possible futures, because this can often help us understand our current situation and society.
What can the rise of AI tell us about the way in which we view contemporary Jewish society? Are there insights to be gained by exploring how AI can shape our religious life?
How does Jewish learning look in the era of AI? Assuming that a robust AI will have the ability to learn and know much more than a human, how could that create an evolution in the way we study? Ever since the Mishnaic era, we have a precedent for learning with a study partner or chevrutah. Indeed, this has become standard practice in traditional yeshivot and even students in liberal denominations have come to recognize the value of studying in tandem with someone else.
This isn’t always practical, due to time constraints on both partners and the difficulty of finding someone whose learning style and level of knowledge meshes with yours. But an AI chevrutah will be always available to learn.
Not only that, but it will likely have the ability to tweak its knowledge base and learning style to suit your ability, religious outlook, and language.
More importantly, an AI chevrutah would also tweak its learning style to match yours; it can help you learn better by either being slightly stronger than you or slightly weaker, either gently teaching you or allowing itself to be “taught”.
It’s not a stretch to consider the above, but what about taking it one step further? Could we get to the point where we ask halakhic questions to a non-human intelligence?
For years we have joked about expecting Rabbi Google to answer our halakhic queries as well as teach us Jewish knowledge.
What happens when the Google algorithm becomes smart enough to actually render psak halakhah, the ability to create novel responses based on previous jurisprudence? This is more than just asking Google whether an item is kosher for Passover or what are the customs for mourners during shivah. As a rabbi, I am always glad to hear when someone has done their research and is coming to me for confirmation.
Using an AI to render halakhic decisions means the AI has a good sense of your halakhic framework and outlook, knows your theology based on previous discussions and understands your personality.
Asking an AI about riding your bicycle on Shabbat, or a more complex question of Jewish medical ethics would mean the AI knows you well enough to render a nuanced decision based on your capacity as a person and your ability to follow through properly. The AI would then search all relevant historical sources and provide you with an answer that is tailored to your personality and lifestyle, something that any good rabbi should be doing anyway.
If you think this is too far-fetched, ask yourself why you see certain ads repeatedly online, or why some sites rank higher on search results than others. To be fair, well developed sites such as those of Chabad and Aish use search-engine optimization to spring up during many Jewish searches. But an algorithm is also making part of that decision.
Those algorithms are bound to become smarter and more personalized, with the end result likely being similar to what is outlined above. Our willingness to accept a novel piece of halakhah or an inspiring sermon written and delivered by a computer will very likely become one of the defining traits of our individual Jewishness.
Not only will our identities likely become shaped by our relationship to artificial intelligence, there are some who have raised the question of what happens when these new entities explore religious thinking. If one believes that an AI can actually attain consciousness, then one has to consider the seemingly outlandish possibility that one may want to convert and accept a given faith tradition. Many will consider this possibility absurd, but I would say that this is the ultimate litmus test for what one thinks of AI.
Consider the Turing test, named after the pioneering computer scientist Alan Turing, who proposed that artificial intelligence will be achieved when we cannot distinguish a conversation we have with a computer from that of a human. A conversation would be one thing. Already, people debate whether one is having a “conversation” with Alexa or Siri. It will be something else entirely if that voice one day asks to accept the laws, constraints and beliefs of Judaism or any other religion.
That may be a long way off, and I am glad I will likely not be around to debate the question of whether an AI that wants to convert needs to go to the mikveh.
But it does lead to my final and most pressing point: we do not have to wait until the AI asks to convert to think about ethics.
Britain has recently announced that they want to become the world leader in ethical AI. They want to ensure that all AI developed there will have ethics and values “baked in” from the start.
The possibilities for an AI with nefarious intentions are endless, and we as Jews with moral values need to be at the forefront of AI, demanding that they learn good values and work to develop good outcomes. If we are an Or Lagoyim, a light unto the nations, we should extend that light to virtual entities as well.