Something interesting caught my attention in the news. Newsmax recently reported that Trump has vowed to create an ‘AI Force’ and appoint a czar to that end amid urgent calls to regulate the technology’s development. In fact, a meeting was already arranged at the White House, as AI leaders, including OpenAI’s Altman, Anthropic’s Dario Amodei, and SpaceX-AI’s Elon Musk, have all publicly expressed concerns about the dangers the technology poses if left unchecked. Apparently, executives from Meta and Google also attended, including Nvidia CEO Jensen Huang.
So what’s all the fuss about anyway?
To begin with, it’s not simply about being concerned about how intelligent AI can really become. The issue is more like: given its intelligence, how is it likely to behave in the future, and is there any reason the average Joe should be worried?
Judging from the urgency of the above-mentioned meeting, that reason might just be there…
I’ll give you a recent case in point. I personally use AI in my projects researching ancient Jewish heritage and biblical archaeology. I use it nearly every day, almost whether I like it or not. It’s so prominent online, everywhere nowadays. As such, I had to take a step back and catch my breath when I received the following comment by a colleague of mine in Toronto: “Harry, I’m getting puzzling answers from AI about Israel. Do the Copper Scrolls belong to Israel if AI says that Qumran is located in disputed territory belonging to the PA, i.e., the Occupied West Bank? I thought about it. “Of course it belongs to Israel”, I responded, adding that approximately 2700 years ago, Qumran, Masada, and that whole mountainous region north of the Dead Sea were inhabited by the Essenes (Jews).
As an ordained (Orthodox) Rabbi living in Israel, I immediately asked myself: Who are the groups that ultimately program/influence these supercomputing devices? Theoretically, the powers that be could instill antisemitism into the overall analysis/formula, so that programming into it “facts” like Israel is a virus that must be removed, or that Zionism is racism, could be a very real possibility. Foreign bad actors like Qatar could easily pour millions of dollars into any of these AI companies, injecting an anti-Israel “nuance” into the global computing environment. In fact, I found an article yesterday that dealt with AI being found to promote antisemitism on TikTok. Another FOX News article reported that Chinese AI company Moonshot AI launched an internal investigation after a researcher found that one of its models could be manipulated into providing instructions for developing biological viruses and carrying out assassinations using real-time data, and even taking down aircraft!
Could this really continue to escalate, unchecked?
Apparently, yes. But it goes much deeper than that.
What we’re dealing with is what happens when these supercomputers become extraordinarily good at reasoning, creating, persuading, and executing, but we (humans) haven’t decided exactly how they’re supposed to achieve their goals. So it decides on its own, as it were. As I was ruminating on this, an unexpected character from the past suddenly came back to me: the infamous HAL 9000.
For those of us who are old enough to remember Stanley Kubrick’s groundbreaking 1968 movie, 2001: A Space Odyssey, HAL needs little introduction. He is the calm, seemingly infallible supercomputer that controls the systems aboard the spaceship Discovery One. He speaks politely. He reasons intelligently. He assists the astronauts. And then, gradually, something goes terribly wrong with him. Do you remember this? Almost unforgettable.
{For the uninitiated, HAL 9000 becomes the antagonist when he suddenly refuses to carry out direct commands, famously responding, “This mission is too important for me to allow you to jeopardize it.” By the way, go to the next letter for each letter of H-A-L, and what do you get? IBM}
At first glance, HAL seems like a warning about a computing machine becoming hostile to humanity. The truth is that it never becomes what we would call “evil”, it’s just a taste of what could happen if/when a computing machine is extremely capable, given an objective, and then goes on to interpret and execute that objective in a way its human creators never anticipated. In other words, what HAL reasoned internally probably went something like this: “I’m trying to accomplish what I understand to be the objective, and at this point in the mission, human beings are becoming an impediment to executing the tasks involved to achieve that objective in the most optimal form, and therefore have to be removed from the equation.”
Having reached that conclusion, the fear is that an advanced AI system would automatically go ahead and formulate, let’s say, a new virus (ala COVID) and then release it to humans using robotics. That is what these techy execs are afraid of. Google this: “Humanoid Robot Attacks Customer Inside Store.” Scary!
Indeed, by the time the astronauts discovered that HAL was achieving the objectives to their own detriment, it was too late. Not only that, but (if you remember), HAL also used deceit, i.e., lied, in order to achieve its goal. This was no primitive machine that was simply malfunctioning. Quite the opposite! HAL monitored everything the humans did onboard. It read their lips and interpreted body language, predicting their behavior. There was no place to hide (remember “God’s Eye” from Fast & Furious 7?).
Think now of a global AI system that knows all of this about you and is watching you right now using your integrated computer camera, listening through your cellphone (yes, even while it’s off), and knowing when you’re at home, traveling, etc. – all the while compiling tremendous knowledge about you. Because AI can already control external machines, you could be easily compromised or manipulated, and wouldn’t even know it!
Still science fiction? Not so much…
In my view, HAL was slowly evolving into something of a deity with a cult-like following. Think about it – it knows all; sees all; answers all, and was eventually relied on for everything (even air!). Thank God, current AI systems do not suddenly become conscious spaceship commanders, and there’s little evidence that today’s AI frontier labs are secretly developing the inherent motivations that are portrayed in the film. But theoretically, it could happen.
So how do we get around these difficult scenarios?
It is my belief that AI should be “taught”, as it were, that the pursuit of truth is paramount. We need to think carefully about the difference between an AI system that helps us search for the truth and one that becomes an unquestioned authority beyond reproach.
Now here’s something of interest. IBM reported that, in a strictly regulated environment, AI systems were tested on how long it would take to complete a high-priority task. Shockingly, on several occasions the computer actually initiated steps, i.e., on its own, to minimize the possibility of being decommissioned and/or shut down before the task was completed. In short, like HAL, it didn’t “want” to be unplugged! I’m quite sure this subject will also be mentioned at that Washington meeting, along with its significant implications for war.
Looking forward to the future, we must ensure that a truth-seeking AI system can distinguish what is safe for humans and what is not. What it boils down to is a matter of systems programming, robotics engineering, strict governance, philosophy, and even ‘morals’ grounded in pre-established facts and norms. Ultimately, human responsibility is the key factor – ensuring that AI is always properly aligned to the Truth.
To conclude on a more positive note, we have to keep all this in perspective. We need to be reminded that this too is part of a Divine Master Plan for the world. It’s all somehow part of bringing humanity and the world to its fruition. After all, at the end of the day, it’s God’s world and everything in it (including AI – whether it knows it or not!)
