While I understand the urge to brush off a technological advancement that potentially causes harm to Jewish thought communities worldwide, I think it best to fully understand the threat, rather than “straw man” it, in order to determine our next steps accordingly. According to Rabbi Mendy Kaminker (“I Asked ChatGPT to Write a Dvar Torah. It Failed,” The Jewish Link, January 19, 2023), the AI lacks the capability to write a dvar Torah that is correct in detail.
At the moment I would say that it depends on how specific the dvar Torah you ask it to write is. But the AI will learn. It will get smarter and better. That is the real fear. And the reason we should not brush it off is that our generation and future generations need to be taught how to spot not only the truth in detail, which the AI will be able to produce, but also the beauty of an original insight, which the AI will have a much harder time doing.
To further advance this point, I would like to analyze the original article and see just how much the AI has learned in less than a week. Rabbi Kaminker makes the point that “Mr. Robot” (his name for ChatGPT) was incorrect when asked to “write an article about Moses translating the Torah into 70 languages.”
While I agree that ChatGPT missed the mark this time around, I asked ChatGPT the same question Sunday morning at around 8:45 a.m., written the exact same way. The AI gave an answer that, while off, was still much closer to home. Mr. Robot, in his answer, had brought an accurate Talmudic story regarding the Torah’s translation, a glimpse into its existence and debate in the world of (I assume secular and non-secular) scholars, and gave a generic although viable point on the vitality of translating the Torah.
Within less than a week, “Mr. Robot” has picked up a lot more knowledge and become much more advanced in scope and fact. This points to the true danger it possesses to Jewish thought, which in the modern era of social media, where anyone can pick up a phone and start typing/recording a dvar Torah, brings with it the likely risk of an abundance of genericism, which will be indistinguishable from AI. he true insight will become a needle in the world’s largest haystack, the internet.
I urge all Jewish leaders to teach and imbue those who they influence with the ability to recognize insight on top of detail and fact. Only once we understand the depths of the Jewish education we are brought up with can we truly recognize insight as opposed to genericism. Only then can we differentiate the human from the machine. Please do not brush off this AI; it is more dangerous than we can imagine.Dovidchai Abramchayev