always fear the toolmaker more than the tool

One of the more interesting facts about ancient human remains buried with tools is sometimes we don’t know if they were tools for domestic use or tools for warfare. In some cases, in many cases in fact, it’s likely they were used as tools for both. An axe could have been used to chop down a small tree, or maybe to cleave away meat from the bone of a fresh kill. Or maybe it was used to kill a rival in a neighboring village. The axe is just a tool to the person wielding it, but a weapon of destruction to the person on the receiving end. What, then, does this have to do with AI?

Artificial Intelligence, better known as AI, has always been a hot topic in science fiction. The idea that we could create another intelligent being would, in some ways, make us gods. Even before the invention of modern machines, humans held myths of bronze statues coming to life and protecting entire islands (such as Talos in Greek mythology). Yet, it’s only in the last decade that we’ve begun to see this as a potential reality. While machine learning has been a thing for a while, it was always marketed as “artificial intelligence” without being artificial intelligence. This has caused quite a stir.

However, far from having created intelligent life and being able to lift ourselves into the pantheon of the gods, we’ve done what our species has always done; we’ve created another tool. 2.6 million years ago we used the Oldowan tools (a stone chipped into a sharp edge to look like an arrowhead) to process whatever we could kill or scavenge. Modern humans, our ancestors, and all offshoots all made and shaped tools. Many primates can use tools (and this isn’t even limited to primates). In fact, among the great apes only bonobos (that I know of) are known to rarely use tools. What makes humans different from the other great apes is our ability to shape tools and manipulate nature to make tools and then use those tools to make other tools.

When we look at what we call artificial intelligence, the reality is while it’s extremely advanced, powerful, and we should be concerned, it’s still in the same category as a 2.6 million year old Oldowan tool. Artificial intelligence, while still beyond the machine learning of the 2010s, still isn’t actually artificial intelligence. It would be better to label it as augmented intelligence; “AI” is a sophisticated tool that augments human intelligence, performing specific tasks, following programming rules and instructions, to produce a result that can sometimes aid humans. It does not, however, exhibit genuine human-like intelligence or understanding because it’s not human and, more importantly, it’s not alive.

In Search of Intelligent Life

What is intelligence?

It’s interesting that in the debate around AI we’ve somewhat failed to answer this question, which is weird because you’d think you need to know what “intelligence” is before defining artificial intelligence. Our failure is not for a lack of trying; we’ve spent thousands of years philosophizing about why we exist and are conscious and aren’t any closer to an answer now than we were thousands of years ago. Hyperbole aside, we do understand intelligence and how it’s gained, or at least to a degree.

One advancement is our understanding of animal psychology and intelligence has led to the realization that they’re not machines. This might seem like a “no duh” statement to us today, but for a while this was the predominant view among naturalists. Research has shown, however, that animals experience the world in ways that suggest consciousness – and this seems to be the case for most animals we encounter (including insects). In fact, that most people wouldn’t question that animals have feelings, emotions, and perceptions shows a modern contradiction for a lot of us meat eaters; if animals are conscious, then there are ethical ramifications to using them as livestock. This is not a debate I’ll get into here, but the point is there is a debate to be had because there’s legitimate concerns about the well-being of animals, because they are conscious.

A lot of animals also exhibit high degrees of intelligence (we’ll define it in a bit). Primates tend to show more intelligence than other mammals (or most other animals), with the Great Apes reigning supreme, even without the addition of humans in the Great Apes family. All the Great Apes show high degrees of cognition – consider that gorillas are likely the “least” intelligent of the Great Apes and yet can engage in sign language and communicate basic thoughts and feelings.

We say animals are intelligent because, in some cases, they show self-awareness. They’re aware of an I, that they exist as an individual and are different from others within their group. This might seem like a small thing, but if we are to consider Descartes profound for “I exist, therefore I am,” the mere act of personal recognition – something human infants can’t do, by the way – then it would apply that for animals to recognize themselves (and therefore have an identity of sorts) would imply some level of intelligence.

Other reasons we’d say animals are intelligent are due to their ability to reason, their ability to learn, to infer things from the environment, to plan, and so on. In the study of human evolution and looking at archeological findings we look for signs of art, critical thinking, reasoning, emotional knowledge, and so on. While all animals are “programmed” to a certain extent, that is from millions of years of evolution, of trial-and-error, being imprinted onto our genetic coding. Likewise, trillions of decisions and accidents went into that, actions made that were random and either worked or didn’t.

For animals, intelligence derives from being conscious. Being part of this world and partaking in the endless trial-and-error that is evolution results in an intelligence that is learned and earned. Animal intelligence (and our intelligence is animal intelligence) has natural problem solving, can adapt to knew information, can infer information, and most importantly can have genuine understanding and comprehension. When we come into contact with information we can compound on our knowledge and continue to process the knowledge as it arrives – this is true of all animals to one degree or another.  

Most importantly, and contrary to modern barbarians, some animal intelligence comes with empathy. Empathy is a sign of high intelligence because it shows (1) you understand that you exist as an individual, (2) that others exist independent from you, and (3) you can change your emotional state to match the state of someone else as a sign of comradery and social building. I’d argue that even in humans who can accomplish great tasks, if they lack empathy then they lack true and proper intelligence. Understanding our place in this universe is the core of anything that matters – everything else is straw. And everyone does have empathy, even those who decry it; they’re just choosing who to have empathy with. Irrespective of what they say and choose, our intelligence developed in part due to empathy (our species evolved in part because of empathy - and empathy is a natural outflow of our evolution).

The thing about empathy is that it’s a choice. This is true among all animals capable of empathy (or who show it) – we’ve observed chimpanzees showing empathy to those within their own clan before going on a raid and murdering chimps in another clan. For humans, we too can choose who we have empathy with – to the point that some people cut off empathy entirely (we call them CEOs). The paradox of high intelligence is that it can’t exist without empathy, but you can use your high intelligence to become an unempathetic asshole and therefore an idiot.

Number 5 Isn’t Alive

One of my favorite movies as a kid was Short Circuit. It came out on VHS around the time I was at the age where watching a movie on repeat ad infinitum is considered perfectly acceptable, and this was one of the movies that I watched. For the uninitiated, the premise is a guy builds a robot (Johnny 5), it’s struck by lightning and becomes sentient, and then acts like a human and gives the famous phrase “Number 5 is alive!” It was his way of saying that he, like the humans who didn’t understand him, was alive.

When Hollywood wants to show that we’re dealing with true, actual artificial intelligence it does one of two things: the robot is hellbent on our destruction, or the robot is empathetic. Movies where AI just wants to destroy us are usually using it as a plot device and haven’t really gone into any deep consideration on the why behind AI wanting to kill us (usually it’s just “Because we suck”). It’s more a commentary on how we view ourselves and our creation than anything else. What’s fascinating, however, is that in movies that want to explore the idea of a machine coming to life, empathy is always the sign of intelligence.

Empathy being the sign isn’t just a trope in Hollywood – empathy is a sign of high intelligence. Artificial Intelligence, as it exists today, not only isn’t empathetic in any true sense, it’s incapable of empathy. In order to have empathy you must have an understanding of who you are as an individual, you must be able to process what life is like for someone who isn’t you. Yet, it’s not just empathy that holds AI back from being proper conscious intelligence.

AI is not autonomous or self-aware. It relies on human input and instructions (you have to create a prompt) and is not capable of independent decision-making or actions outside of its programming. If given a task it will follow that task, but it won’t sit there and go, “You know, I should really do this task even though no one has asked me.” Now, in some jobs that might actually be a good thing, but it also shows that AI isn’t “intelligent” in the proper sense of the word.

Anyone who has used AI in a proper context will tell you that AI is absolutely horrible at having true understanding and contextual understanding. An example of this would be as follows:

Imagine you need to build a website that will sells something. You need to create a button that says “Buy.” AI can absolutely build that button with no problem. But as the repository grows and you get into the hundreds of files and the thousands of lines of code, AI simply will lose context of what is needed. Even when focusing on a single problem it might forget or ignore the broader context of the problem, something humans aren’t apt to do.

Our knowledge compounds and we form connections between what we learn. Sometimes we can learn something that will trigger a connection to some far-fetched thing, but suddenly we can make that connection. AI’s knowledge doesn’t compound, but instead is fragmented to what is needed in that moment. Now, one of the reasons for that is due to a limitation in processing power. One could argue that if the processing power were higher, this wouldn’t be an issue. But the processing power needed to have an AI system that could build a complex web application with little human involvement simply doesn’t exist.

Even if the processing power existed, AI as it currently stands still wouldn’t compound knowledge in the same way that we do, nor would it process knowledge in the same way that we do. The way LLM’s work is they’re predictive, not intelligent. They revolve around complex data analysis algorithms rather than understanding what something is – they can recognize objects in an image but won’t know why those objects matter or understand what they truly are. Because of this, they struggle immensely in situations requiring common sense, abstract reasoning, or creativity. The more mundane and repetitive a task, or the more general a task, the better AI is equipped to handle it.

The Good uses of AI

Since AI isn’t some other intelligent being and is just a complex data program, it’s safe to call it a tool and when used properly it can result in some amazing things. I use AI everyday for my job – and it’s accelerated my learning and my abilities by years. I’m literally years ahead in one of my skillsets thanks to AI and using it to learn. That skill set is not history, research, philosophy, or any other liberal art. When it comes to crunching numbers, AI is great. When it comes to telling you why those numbers matter, leave that to the humans.

On a scale bigger than me, AI has been used to diagnose diseases, help doctors triage patients more effectively, help increase the turnaround time on cancer screenings, and the list goes on in healthcare alone. It’s helping in office productivity, in computer science, engineering, and a host of other fields. The hard sciences are poised to undergo a revolution thanks to AI.

For students with learning disabilities, AI offers assistance either by helping them reinforce positive behaviors and actions, or going so far as to act as a learning aid. Overall, AI is a major gamechanger for students with learning disabilities, opening up equitable access to services given to students without disabilities. It’s not just education either – Microsoft has a list of people who use Copilot to remove barriers to accessibility. As someone with moderate ADHD and severe anxiety (coupled with depression), I’ve used AI in some emergency situations to work through an anxiety attack. What’s crazy is that it’s worked and has even helped me organize my days better.

Like all tools, AI has its use and when used properly – as something to aid humans – it helps us achieve great things. AI has the potential to be one of, if not the most powerful tool we’ve ever created as a species. If used properly, it could advance us centuries in a matter of decades. I don’t mean flying cars or living on the moon, but I do mean significantly longer lifespans, higher quality of lives, the eradication of multiple diseases, and so on. Far from a utopia, we’d still have problems, but life could be significantly better.

If

We come across the remains of a human from 9,000 years ago and see a spear next to him. We wonder – did he use that spear to hunt? Was it decorative? Was it just a ritualistic item (this would be the first guess of any archeologist, even if just for posterity’s sake)? Did he use it in battle? How did he use this tool?

If AI is used properly, then…If we use AI for good, then…If. Is there a more hopeful and hopeless word in the English language? Describe an emotional state of having absolutely no hope while simultaneously holding absolute hope. German accuracy would demand a 12-syllable word. English pragmatism has condensed it to “if.” AI is powerful, but it is not good. It’s also not bad. It’s just a powerful tool, and like any tool what matters is how it’s used.

We already know that AI itself is problematic. From energy consumption to copyright laws, to its use in unnecessarily replacing workers, it’s not difficult to make the case against AI. The reality is that its use is likely going to get worse as the years go on and AI becomes more powerful. Like any other powerful tool created by humans, there will be those who want to use it for good and those who want to use it for themselves.

Look at our ability to split the atom. As a tool, nuclear energy provides us a surefire way out of our climate crisis. Nuclear energy could create a world of cheap electricity, which would allow for more development without the environmental costs. Yet, we also know that if people take shortcuts to make a profit and save some money, that clean energy could turn into an unmitigated disaster. While there are multiple checks on our nuclear power plants today and the risk of a meltdown is low, do you really want to take those odds in the era of Musk chainsawing the entirety of the United States?

Likewise, nuclear energy comes with the risk of nuclear weapons. While difficult, it’s not impossible to take the leap from “we have nuclear power plants” to “we have nuclear weapons.” Nuclear weapons run the risk of completely wiping out our species, or at least setting us back a few thousand years for a few thousand years.

Any tool we create, no matter how well-intentioned, no matter how helpful it can be to our species, can also be used for harm. AI is no different.

Don’t fear the toolmaker – become the toolwielder

I don’t fear AI – it’s a tool that can be used – but I absolutely fear the men in power who will use AI for their own advantages. I fear their use of AI, or misuse of AI, in scenarios where this tool shouldn’t be used. We live in an age where we have to ask drastic questions, one of which is this – how difficult would it be to create AI to run an execution machine? If you can remove the human element from pulling the trigger, it makes it much easier to send people to their deaths. Put in more drastic terms, what if the Nazis had AI in their camps – a sickening thought, but is it far-fetched today?

What this means is that we need regulations around AI and its use across all spectrums of society. However, considering the current political climate, hoping for that – or for regulations that are good, meaningful, and helpful – is akin to hoping that aliens will arrive and offer to fix everything for us. We can’t regulate what we don’t understand, and the average American doesn’t understand AI while the average congressperson can either benefit from the corruption around AI or thinks the internet is a series of tubes.

 I don’t know what the solution is – I do know that the answer isn’t to avoid AI. While there are privacy concerns and legitimate reasons to avoid using it due to energy consumption, we also must accept that it’s a powerful tool that’s helpful to millions of people and therefore is here to stay. Thus, the best way to make sure a tool is used properly is to understand and use the tool yourself. While you can’t do this in every case (like nuclear energy), you can do it in the case of AI.

The only way forward that I can see is to accept the reality that AI is here to stay, so it’s a tool I will use and that I will use for good, or as best as I can. The reality is, people who can use AI efficiently are going to get promoted and pushed up – so long as there’s an economy that can support jobs AI can help with (no longer a guarantee!), those who understand AI will be the ones getting those jobs. It would make sense then that since AI is a tool we should want good people using it.

I think back to all those tools we’ve found alongside humans and wonder how they used them, but ultimately it boils down to that individual and how they chose to use those tools. The same is true for us and AI – it’s our very advanced stone tool. We can run away from that and let the barbarians use the tool, or we can begin to learn it ourselves and maybe, over time, those who have concern for humanity will be the ones dictating how AI is used.

If people who hate the marketing around AI and all the proponents, but who have a heart for humanity, were to overcome their fears and doubts and use this tool better than those who only have a heart for themselves, what good could we do in this world? If we learn this tool, if we use this tool properly, if we wrestle it back from those who would only use it to advance themselves and protect their positions, what could we accomplish?

If.