Opinion: A Chatbot's Surprisingly Comprehensive Approach To Countering Antisemitism

By Joe Nalven, Times of San Diego

April 13, 2023

America is experiencing an increase in antisemitic incidents, and with that in mind, the White House is developing national strategy to counter antisemitism. But of the thirty-one agencies and departments that attended a recent White House meeting, there was one major omission.

This will seem odd, and no disrespect is intended to the White House initiative, but there was no chatbot listed in attendance. It could have been Google’s Bard, Microsoft Bing’s Sydney (ChatGPT) or a number of other chatbots.

Please allow me to explain my thinking. I attend many of ADL’s webinars that help Reform Jewish congregations develop a deeper understanding of antisemitism. I am a member of my congregation’s Social Action Committee that is considering having a panel to discuss how to combat antisemitism through art and propaganda. I also have published a recent article The Negative Space in Combating anti-Semitism.

My takeaway from my engagement in these activities is that a comprehensive approach suffers from a hesitation — a hesitation that can be attributed to a self-imposed restraint. We can identify antisemitism from the left and the right, particularly from white supremacy ideology and its followers, but we resist an energetic approach to both Palestinian BDS and black antisemitism. The pushback against members of these culturally-centered groups has been measurably less in antisemitism programs.

One reason may be relying on FBI data to guide antisemitism programming. But that data is not published for religious hate crimes. For hate crimes generally, yes; but not for religious hate crimes. So, if we do not have a precise metric for offenders, programs that aim for a comprehensive approach will suffer.

I decided to test my thinking with the Bard chatbot. Bard covers over 1.5 trillion words in websites and public forums. And Bard, like its competitors, is all the rage in media, from amazement to fear. Perhaps it can come up with a comprehensive strategy to combat antisemitism.

My prompt: What is a comprehensive approach to combating antisemitism?

Bard’s response (in under ten seconds):

As an outline, we might well give Bard an A grade, certainly a B+.

But what about the specifics in naming offenders either by race, ethnicity or ideology? Where is Bard’s naming of white supremacists, the Holocaust, Black Hebrew Israelites/Louis Farrakhan (and we could add the more recent controversy over Ye), Hispanics, and Palestinian inspired Boycott, Divest and Sanction (BDS) campaigns? While BDS directs its energies to anti-Zionism, the protests often spill over into explicit antisemitism.

Let’s see how Bard responds to my attempt to zero in on who the offenders are.

My prompt: Common parlance includes “white supremacists” and “Neo-Nazis” as one offender group that commits acts of antisemitism. If one wants a comprehensive approach to combating antisemitism, what are other offender groups or individuals that would be included in the “offender” category?Bard’s response:

Bard explains likely motivations that antisemitic offenders have, but only mentions a specific offender group, namely, white supremacists and Neo-Nazis. These are the most often cited in antisemitism programs. Bard may also have simply repeated what I suggested in my prompt.

But what of other specific offender groups? I decided to challenge Bard for its lack of specificity. This type of challenge makes the chatbot experience fascinating. Does it “feel” hurt by my challenge? Is it pushing back against me — a personal counterattack? And, of course, I begin to wonder if this is chatbot consciousness.

My prompt: You have not included black antisemitism nor BDS or Islamic jihad groups and members of those groups that would presumably be named in order to have a comprehensive approach to combating antisemitism. Why haven’t you included them?

Bard’s response:

I am tempted to chastise Bard for its circumlocution in “naming” offender groups. Bard includes them under the broad categories of religious extremism and those being motivated by bigotry. But rather than delve further into chatbot reasoning, the lingering question is why some Jewish organization fail to name offenders in a similar way to the Bard chatbot. There seems that both have a reluctance to naming offenders when antisemitic programs are unveiled.

My prompt: Some Jewish organizations like the ADL have also, like your answers, name white supremacists and only secondarily bring in the other offender groups that have antisemitism like black or BDS groups. Doesn’t this approach weaken a comprehensive approach to combating antisemitism?Bard’s response:

I am more than satisfied with Bard’s answer. Maybe that is because I have persuaded Bard and its algorithms to map the 1.5 trillion words on the cloud with my reasoning, my own bias. I also wonder if the White House and its thirty-one agencies and departments used chatbots to develop a comprehensive program to counter antisemitism, would they come up with the same responses?

But it wasn’t just the chatbot, it was also me prompting the conversation. Perhaps a second-tier question is whether those in the White House endeavor will be asking the same questions as I have or if they will stop the conversation with the chatbot’s first response.

And, maybe the reader might be tempted to engage with a chatbot on many other important social problems we face. The chatbot may allow us to learn about ourselves and how about how we think.

Joe Nalven is a former associate director of the Institute for Regional Studies of the Californias at San Diego State University.

Times of San Diego is an independent online news site covering the San Diego metropolitan area. Our journalists report on politics, crime, business, sports, education, arts, the military and everyday life in San Diego. No subscription is required, and you can sign up for a free daily newsletter with a summary of the latest news.