AI, AI, AI (Yiddish for ‘Trouble Ahead’): Depends On Whom You Ask

The “news” is almost always about what’s happening “today.” Yet, when historians and the public look back years later, this distancing perspective offers a completely different picture. So while right now Israelis are into the upcoming elections (as are Americans in the US), years hence the main development of the mid-2020s will be something else entirely: Artificial Intelligence. Which doesn’t mean that today we know how this will turn out; quite the opposite! Here’s why.

Like most revolutionary technologies, AI can be different things for different people or sections of society – and it’s not always clear at the start who wins and who loses. Take the Gutenberg print revolution for example. The Church thought that print would be terrific as then relatively inexpensive, printed prayer books could be distributed to the masses, enabling more people to participate in church services. But it didn’t take long (about 70 years later) for Martin Luther to exploit print and widely disseminate his theses through posters on church doors to start the Protestant Revolution that split Christianity in two (or more).

So do take my analysis here with a grain of salt or two. Nevertheless, once in my academic career I succeeded to be well ahead of the game: 45 years ago in 1981 I published the first-ever academic article on the legal ramifications of AI harm (https://www.researchgate.net/publication/229067941_Frankstein_unbound_Towards_a_legal_definition_of_artificial_intelligence).

Who wins and who loses with AI? At this stage it can be stated as a generalization that AI is very popular on the macro level (governments), becoming highly unpopular on the micro level (most workers), and somewhere in between these two un/popularities on the meso-society level (corporations).

Governments see AI as a tremendous economic and security booster – able to provide the same services (or more) to the public with far less cost. Moreover, as in every other important technology, there’s the fear that “if we don’t do it, our enemies will get ahead of us.” Thus, for example, the Trump Administration is trying to ward off what it sees as over-regulation of AI by various states, so that AI companies can be given virtually free rein to “do their thing” at breakneck speed. Of course, “breakneck” speed can turn out to be literally true – breaking the neck of society.

That’s precisely why the general public in the West has become increasingly wary, even fearful, of the AI boom. There are several good reasons for such trepidation. First, there’s the worry that AI’s very success could lead to massive unemployment as employers use it to do more work with less wo/manpower. AI doesn’t have to be “better” than humans; it’s sufficient for AI to be good enough: cheap, quick, and widely deployable – Artificial Good-Enough Intelligence.

Second, AI has been found at times to provide bad advice to people seeking psychological help – even encouraging suicide on occasion! This isn’t absurd; human beings have a penchant for what’s called “cognitive surrender.” A recent study found that participants followed AI’s advice 80% of the time when given the wrong answer! (https://futurism.com/artificial-intelligence/study-do-what-chatgpt-tells-us).

Third, AI has exponentially increased the ability of “bad actors” (e.g., criminals) to do harm, whether by massive “phishing” expeditions or by more easily breaking security codes and other cyber self-defense measures.

Fourth, and certainly not least, AI uses massive amounts of electricity through its humongous data centers, potentially driving up electricity prices for everyone. NIMBY (Not In My Back Yard) has reemerged bigtime in America.

As a result of these countervailing trends, AI companies are caught in the middle – not wanting to fall behind their competitors but also not wishing to alienate their customers and employees. As a result, in Europe as well as in the U.S. a most unusual phenomenon has emerged: the AI companies themselves are asking governments to pass AI regulatory measures so that they know what they can and cannot do with their AI algorithms. As for the electricity bill issue, AI companies are so worried about this public backlash that they’ve begun serious thinking about placing data centers and their power sources on… the moon!

All this involves potential direct harm (and benefits). However, more subtle but no less dramatic is AI indirect harm. As we increasingly rely on AI to do our heavy cognitive work (after all, AI can take into account hundreds of thousands of factors – something impossible for humans to do), there’s the danger of lowering human innovation, forgetting (or never bothering to learn) stored knowledge in our brains, and even losing the ability to think critically (both meanings).

A second type of indirect harm would piggyback on a problem already evident before the AI era: non-factual thinking and argumentation. AI is capable of claiming things that don’t exist (“hallucinations”) and/or displaying “reality” better than it really is (e.g., misleading advertising, deceptive self-presentation, etc.).

Where’s Israel in all this? The picture here too is mixed. On the one hand, Israel is a mini-powerhouse in artificial intelligence, with over 2,300 active AI companies. However, the country and its people have had far more pressing things to worry about these past few years than AI-driven un/employment. Thus, it lags greatly in any public debate on the issues involved, not to mention Israel’s government and regulatory apparatus that seems to be “missing in action” on this score.

So, if you have a chance to talk to anyone running for Knesset, ask them what’s their position on AI. They’ll most probably answer: “let me get back to you on that” – and then run off to ask Gemini/ChatGPT/Claude/Perplexity et al for guidance…

At some point, Israel’s leadership will have to address the AI revolution (tsunami?) as it starts making serious inroads in society and the country’s economy. Who will be Israel’s (and the world’s) 21st century Martin Luther? We will soon find out – probably for better and for worse.