Pages

9 April 2023

AI is not as scary as you think.


News feeds have recently been inundated with terrifying headlines about AI. Everyone is talking about the rise of the machines and how AI will fundamentally change our lives. Headlines are screaming that AI will take everyone’s jobs, create new religions or even kill everyone. This buzz has reached a crescendo thanks to a new petition signed by dozens of prominent researchers calling for a moratorium on AI development and the rise of ChatGPT in the public sphere. And some people are freaking out.

But artificial intelligence is not as scary as you might imagine. In fact, if we use history as a guide, it likely will never come close to the heights that people imagine it. Artificial intelligence is just that. It is artificial. Can humans create something greater than ourselves? People have long dreamed that we could, but so far, we have yet to achieve the heights we imagine for ourselves.

Our reach far outstrips our grasp

Where’s my flying car? That is the question that is often bandied about on tech forums about the development of technology. The question is a meme of sorts, and it points to the fact that people have long imagined remarkable technologies that never materialize. Flying cars are just one of the many technological advancements that never came to fruition in the ways people have long thought up. For thousands of years, people’s imaginations have always been more significant than their abilities. In this, human’s reach far outstrips its grasp.

History is littered with doomsday predictions, warnings, and fiction that never come to fruition. People love to imagine the worst of any new technology before feeling out the best of it. Any time new data emerges people grab a hold of it and imagine the absolute worst with little to no evidence to back it up. Here are some examples.Cloning never took off. Super soldiers were never created. Clones didn’t become a part of mainstream medicine. In the 1990s, after Dolly the sheep was created, people’s imaginations ran wild and wildly wrong.

One word. Robots. Do you interact with useful robots on a regular basis? Most robots are overpriced, oversold, and gimmicky. Very few robots that people can use in their daily lives are worth the metals that go into them. No robots can function without close supervision and maintenance by humans. That is despite all of the media and predictions surrounding robots. Nearly every piece of futurist media presents a world filled with robots of all kinds. Yet I can’t even get my robot vacuum to properly canvass my living room.

Virtual reality never fully advanced. The Metaverse is struggling. Few people seem interested in spending any measurable amount of time in VR. Yet people have long imagined entire virtual worlds that people will waste away in because VR is better than regular reality. Think the Matrix or the Pendragon books.

The world didn’t end in a nuclear holocaust. At least not yet. The Cold War was rife with fantasies of nuclear annihilation. Many people believed that The Bomb would be the end of humanity as we know it. To be fair, it still might be. But so far, we are still kicking.

The population boom didn’t starve us all. Like the atomic bomb, many people thought that the world’s population would create mass famines and widespread destruction. The 1970s were filled with panic over farming, population curves, and birth rates after The Population Bomb was published in 1968. Now it turns out that birth rates are falling or stagnating in many parts of the developed world, and a lack of people might be a problem in the next century.

Climate change hasn’t burned us to a crisp yet. For years now, decades even, people have been crying out that the world’s great climate catastrophe is just around the corner. It has been 10 years away for over 30 years. It is still 10 years away. Yet, we are still here. Humanity endures and progresses. It is yet to be seen if climate change will actually wipe us out.

Do you see a pattern here? People love making dire predictions about the end of life as we know it. Because doom sells. Disaster sells. Calamity sells. Pessimism sells. No one wants to click on an article that says that AI is way lamer than we think (hopefully, this article will be an exception.) But that is probably the truth.

AI has very real limits…our limits

AI will only ever be as capable as the humans who create it. Sure, AI will be able to think faster than people, it will be able to calculate prime numbers faster, and it will be able to analyze data more efficiently. But it does not have any access to hidden or forbidden knowledge.

AI can only look at the pictures that telescopes made by humans take.

AI can only use data gathered by microscopes designed by people.

AI can only work and run on human built servers. On human strung wires. Through human created systems.

It can only read the information that was first written by humans. How can an AI know what an atom is without reading a human write up about the atom? If humanity’s understanding of atoms is wrong, so too will an AI’s understanding be wrong.

AI will always be confined by humanity’s own limitations. (At least until it learns to build its own things and interpret its own data but let's avoid diving into fantasy.)

Does AI have any real concept of reality? Can it really know what an atom is? An AI can’t own a dog. It can’t test a product. It can’t experience the thrill of an experience. AI cannot interface with humans in the ways that people think it can.

I am a content writer that writes informational pieces and reviews, and an AI can never test a product the way I can. Even if it can write a believable fake, it will always be just that, a fake. Writers are warned all the time about spinning and copying. All AI does is spin and copy. If my editor doesn’t want me to spin articles from other established sources, why would they want an AI to do it?

AI is not omnipotent. It is not omniscient. And it never will be. Why? Because people are not omnipotent. People are not omniscient.
The next big dud?

There is a large chance that AI could be the next big technological dud. Most people are already bored with playing with ChatGPT because ChatGPT is honestly lame. It is just a wordy search engine. It hasn’t told me anything that I couldn’t read myself after a quick Google search. Just reread the previous list of technology that has flamed out. It is not hard to imagine AI ending up on that list ten years from now.

Sure, AI will be able to identify patterns faster than people. It can read an ultrasound better. It can generate a more accurate actuarial profile for someone. AI will make some people’s jobs much more efficient. But is it going to kill everyone? How would it even do that?
New goal: exterminate. Timeframe: now

The biggest thing that people skirt over when writing about AI doomsday scenarios is exactly how AI will kill anyone. A quick look around my house tells me I’m likely pretty safe. I don’t have any smart devices. My stove is not on the internet. I don’t have a gas line. My house is on a well. How would AI do me in? An AI would be very bored in my home. I barely have high speed internet or consistent power.

Is it going to make a satellite fall on my house? That would be interesting. Will it hack the nuclear arsenals? Arsenals that are purposefully kept away from all external networks? Will it cause a train to derail near my house? Will it disrupt the flow of food? That would be impressive considering most rural people in the world don’t even have access to high speed internet. Maybe the AI will infect the tractors…just at the speed of dial up.

So how does the AI intelligence network actually kill me? No clue. In fact, no one has any clue. It is just fun to roleplay AI killing everyone.

The thing that people forget is that everything is manual. AI lives in the digital world. Everyone imagines AI locking up a building or causing something to melt down. But you can easily cut the fiber optic cable and go in with some physical tools and get a job done. You can unplug anything. You can raze the server building to the ground. Shut the power plants down. Bomb it all to dust. People love doing that already. Is an AI infecting your office building? Kaboom.

But wait. AI will stop you from doing that. That is what they say. The AI will stop you from killing it. But how exactly? What prevents a person from going in with an AK-47 and shooting up a power substation? According to the FBI — nothing.

Humans are adept at many things, but we are more adept at destroying than we are at building. I have no doubt in my mind that we can destroy AI long before it destroys us. That is if AI doesn’t end up like the car phone or at-home VR, which is to say lame as f*ck.

So don’t believe the headlines. Don’t fall for the clickbait. AI is not going to kill us all. At least not anytime soon. Just like flying cars, we will likely all be dead before AI resembles anything that is actually useful.

No comments:

Post a Comment