In June 2023, 25 renowned Haredi rabbis signed a proclamation publicly banning the use of artificial intelligence chatbots. They asserted that the information such chatbots can generate is rife with heresy and abomination and that exposure to them will inevitably result in “falling into a minefield of danger.” This ban follows others promulgated years ago against use of the internet, then a new technological innovation deemed unacceptable for the ultra-Orthodox community. Yesterday’s internet has become today’s ChatGPT.
Are these rabbis on to something? Certainly, the challenges that new technologies pose to observant Jewish communities are manifold, and perhaps AI really is different, as we are often told — perhaps we may finally have created, or may be inexorably creating, a golem that will ultimately control us. I have my doubts and proceed on the basis that AI won’t transform human life out of all moral recognition. Jews have been confronting threats from technology for as long as there have been Jews and technology.
It is how technology is used that matters. History would suggest that rather than resist steel because it can be beaten into swords, we do better to embrace it and make plowshares instead — not least because we as a species have never proven able to choose not to create something that has been in our power to create.
If I am unfazed by the challenges of today’s technology, it is because this is not the first or even the second time that observant Jews have reacted with alarm to new technological developments.
More than 400 years ago, the invention of movable type engendered a great deal of consternation among Jewish authorities. My review of the historical record suggests that they had six major concerns: quality control, permanent error, the damaging of reputations, wasting time, the diminution of authority, and the easy accessibility of inappropriate material.
Quality control. The 17th-century sage Rabbi Yosef Shlomo Delmedigo noted that manuscripts were extremely costly to write — and costly to copy. As a result, relatively few manuscripts were produced, and only those of genuine merit survived. In an elegant play on the words of Esther 8:17, Delmedigo complained that with the printing revolution, rabim me’amei haarez mityaharim — “many ignorant people become boastful”—and, in a desire to become famous, “make crooked that which is straight.”
Permanent error. The famous 16th- and 17th-century Talmudist Rabbi Shmuel Eidels, known as the Maharsha, noted the impossibility of rectifying an error, innocent or otherwise, once it was replicated many times in printed works. He pointed out that some who did not understand a passage in the Gemara, Rashi’s commentary, or Tosafot — the most fundamental of Jewish texts after the Bible, along with its two most important commentaries — “corrected” what they assumed to be a corrupted text in the margin of their copy, upon which a printer, thinking the correction authoritative, substituted it for the original. Absent a later correction from a scholar with access to the original text, the error would remain forever.
The damaging of reputations. In 1619, the responsa collection of Rabbi Meir of Lublin was published in Venice. It contained disparaging remarks about one man involved in an acrimonious dispute in Mantua. His aggrieved children prevailed in an appeal to the rabbinic authorities, who ruled that the relevant page be reprinted without the offending passage and that anyone who owned a copy should have the offending page replaced with the new version. The dangerous, long-lasting power of the printed word was recognized and addressed — but only with very considerable effort.
Wasting time. A 1587 Ferrara enactment bemoaned the fact that the easy accessibility of mediocre books pushed worthwhile ones, “full of wisdom and knowledge,” to the margins. It would now be possible for people to spend their time in frivolous pursuits rather than with books that could bring them much benefit.
The diminution of authority. The much wider availability of printed texts undermined the hitherto exalted status of the learned elite who had, until that point, enjoyed a virtually exclusive monopoly on knowledge and its dissemination. Study with a recognized authority in a yeshiva, the traditional source of Torah knowledge, suffered a collapse as a new community of learners was created of all those who had access to printed books. (The problem was serious enough that Sultan Bayazid II, the most powerful Muslim ruler at the end of the 15th century, outlawed printing altogether.)
The easy accessibility of inappropriate material. In 1529, some 30 years after the establishment of the first printing press in Salonika, the city’s rabbinic leadership forbade Jews to print anything without the permission of six rabbis. They regarded stopping the activity of printers who “published a number of things that were not appropriate to print” as so important that transgressors were to be banned.
Despite all this, of course, printing flourished. This was in part because many in the Jewish community recognized how useful it was. Of Gutenberg’s invention of printing, Rabbi David Ganz (1541–1613) wrote in the entry for the year 1440 in his historical Zemah David that “nothing as valuable as it is found in all the wisdoms and clever devices from the day that God created man on the earth,” and that, “were it not for printing, God forbid, Torah would have been forgotten from Israel.” Printing created what Benedict Anderson has called an “imagined community” of learners. More people were able to learn more deeply than ever before in Jewish history.
For a religion so thoroughly devoted to textual study, the internet, which proliferated the production and distribution of text vastly beyond anything we had seen before, was a world-shaking development even beyond Gutenberg. Not surprisingly, it was greeted with even greater alarm. In January 2000, a ban (not the only one) was promulgated by 29 Haredi rabbis, “declaring the internet to be the greatest menace ever to face Jewish culture.”
Once again, all the concerns raised by the arrival of printing reappeared. If quality control suffered with the arrival of mechanical reproduction, it may be said to have disappeared entirely with the growth of the internet. Suddenly, you no longer needed to persuade a traditional newspaper, magazine, or book publisher that you had something worth saying. The range of publications on the internet quickly became so broad that almost anyone willing to offer his copy without charge was likely to find a publisher somewhere. Today, you can also publish and sell your own book through Amazon and your articles through Substack, and share your thoughts on any topic in long or short form on Facebook or X, as Twitter is now called. You can have your article up one minute after you have finished it. Almost no financial resources are necessary, and not even the tiniest measure of professional or scholarly expertise is required. As Clay Shirky, a student of the social and economic effects of internet technologies, notes in Cognitive Surplus, “the easier it is for the average person to publish, the more average what gets published becomes.”
As for the proliferation of error, innocent or otherwise, the internet is forever. In a 2010 New York Times Magazine article aptly entitled, “The Web Means the End of Forgetting,” Jeffrey Rosen wrote, “the internet records everything and forgets nothing. . . . Every online photo, status update, Twitter post and blog entry by and about us can be stored forever.”
As for the ease with which one may damage reputations (one’s own included), the story of the community that reprinted a single page of Rabbi Meir’s responsa and inserted it into existing copies of the book it came from now seems quaint. The terrible effects of speaking ill of another, lashon hara, have received enormous attention, because in Judaism (although not in Judaism alone), words have always been recognized as having enormous power. Today, bloggers, vloggers, Tweeters or Xers, Facebookers, Instagrammers, TikTokers and the rest can hide behind pseudonyms as they destroy a reputation with a click — far more broadly than was ever possible before, as physical communities are replaced by virtual ones, and much of the globe can be reached in an instant. We have become too habituated to gossip, so it is useful to be reminded that the Talmud (Bava Metzia 58b) observes that “one who shames another in public is as if he is committing murder.”
And of course, the internet is a colossal temptation to waste time. A recent survey concluded that the average U.S. teen spends close to nine hours a day in front of a screen — with more of it devoted to gaming, social media, and surfing the Web even than to watching TV or videos. Less than half an hour is spent on e-reading or creating “content.” Low culture predominates at the expense of exposure to worthwhile information.
Finally, the rabbis noted of the rise of printing that suddenly anyone could be an expert on anything and that it was too easy to access inappropriate material. No one would dispute that there are vastly greater problems today than the ones that printing created, but they can be examined through the same framework.
When it comes to quality control, clients routinely walk into their lawyer’s or doctor’s office claiming expertise based on internet research. The challenge to Judaism is perhaps greater. Rabbi Hershel Schachter of Yeshiva University once pointed out: “Who needs a rabbi or rebbe to deliver a judgment about laws. . . . Anyone who studied in a yeshiva can deliver a judgment or adjudicate on the basis of his own reasoning. . . . It is possible for everyone to consider himself a scholar and halakhic adjudicator and arbiter even on weighty matters as if he knows all of the sources and all the opinions on his own.” Now that these sources and opinions are widely available and searchable online, one does not even need the yeshiva.
As for the accessibility of inappropriate material, what is available on the internet still has the power to shock most of us. Pornography of every kind is just a click or two away. When I was young, adolescents (let alone children) who wanted to get hold of inappropriate material had to interact in a public space with an adult bound by law to prevent them from achieving their goal. Even were the adult willing to sell the offending publication, the need to speak one’s desires aloud presented a very embarrassing barrier that might itself have short-circuited the transaction. Today, my grandchildren can, if they wish, access all kinds of material I would not even have been able to find half a century ago, let alone acquire. A similar effect is created by the shift from the telephone in the hall, when everyone knew whom you were talking to, for how long, and often about what, to digital platforms of who knows what kind, populated with who knows whom, behind closed doors — for hours.
So the internet presented all the same challenges introduced by printing, multiplied a thousand times. And yet, with the exception of Haredi communities, we — that is, the Modern Orthodox, those belonging to more liberal denominations, and those identifying with no denomination at all — made our peace with the internet. Serious though its negative effects have been, we are obviously enjoying its remarkable blessings.
Certainly the internet has made it possible for Jewish learning to flourish on a scale hitherto unimaginable. The number of websites and apps with Jewish content available literally at one’s fingertips is staggering. The easy access they provide for rabbis, teachers, students, and any Jew with any question at any time has revolutionized Jewish study and Jewish life, greatly broadening the community of learners and bringing Jewish practice to many who would otherwise have had difficulty accessing the tradition.
What, then, of artificial intelligence and early broad-access apps such as ChatGPT? First, it is clear that, just like the internet, they are here to stay. AI may seem new, but it has been in development for the better part of a century and in use in large organizations for decades. Rather than seek to throttle the exposure of the Jewish community to AI, which will surely prove futile outside the Haredi community and perhaps ultimately inside it, too, communities need to think hard about how to maximize its advantages and minimize its threats. These are choices. We routinely teach our children and students how to make choices: who their friends should be, what to look at and away from when walking in the street, what books to read, what foods to eat, and much more. The options available in each of these cases are many, and some are destructive, and our responsibility is to educate the next generation to choose those that will enhance their lives. The choices regarding technology are no different.
I believe that artificial intelligence — much like the internet and, before that, printing — will provide more benefits than harms. AI is likely to transform teaching, and Jewish education with it, by making the educational experience uniquely responsive to each student’s knowledge, ability, and preferred way of learning; by increasing access to the right resources; and by tailoring the choice of subject matter to engage the student as effectively as possible. AI is also likely to transform the world of learning as much as the internet did, and perhaps more. Might it dilute the human relationships crucial to good teaching, or deepen the “digital divide” between rich and poor? Might it isolate as much as it connects, just as social media appear to be doing today? Of course. But these are outcomes we can influence by the choices we make as a community. AI is here to stay, whether we like it or not. The Jewish community learned to live with every technology developed until now, and we will do the same with today’s technology, too.
If I have one hesitation, it is that new technologies cannot always be adequately described through analogy. For instance, one thing fundamentally different about the internet from the world of books is that it is interactive — and insistently so. I noted before that the internet is forever; but it’s not just forever, it’s forever 24/7. The most painful example of this is the effect of social media on young people today. Whatever its upside, it’s impossible to escape. Once upon a time, a child bullied in school could get some relief each day when she came home. Today, the onslaught is nonstop, with tragic results common enough that the term “cyberbullicide” returns 150,000 results on Google. Meanwhile, the majority of young people (particularly young women) who have the good fortune not to be bullied are nonetheless spending hours online day and night looking despairingly at photoshopped images that present impossible “models” to which they are encouraged to aspire. In their hyperconnected spaces, they’re more isolated than ever before. It’s a commonplace that social media have helped those of us in the richest countries in the world raise a group of teenagers who are more anxious and depressed than any generation in memory.
So, just as the internet has proved damaging in a way different from any technology before it, it is of course possible that the potentialities of AI are so great that we simply do not and perhaps cannot know what we are letting ourselves in for. I am not sure anyone knows what artificial intelligence might make possible. We are, to some extent, in uncharted territory.
But this, too, has happened before. In the early 1940s, the United States created the first fission bomb. Many were the predictions of utter global disaster — not unreasonable predictions, all things considered. And yet, in 80 years, the fission bomb has been used only twice, to end a world war that threatened millions of additional casualties had it not been used. Its vastly more powerful successor, the fusion bomb, has never been used at all. I am optimistic that we will find a way to make the right choices, or at least avoid the worst wrong ones, with AI, too, as well as with all the technology at our fingertips.