History is Clear: AI Does Not Need Regulation!

‍

‍

Lots of people are calling for the regulation of AI. but frankly, those people don’t understand technology. Someone who does is Sam Altman who argued, “It’s a hard message to explain to people that current models are fine. We don't need heavy regulation here. Probably not even for the next couple of generations.” By “hard message” he means you’re not smart tech guys like us and can’t understand what us smart people do know, so stop worrying your low-IQ heads about it. Now I’m not going to give you the technical reasons why AI shouldn’t be regulated because, well, apparently, you’re not smart enough, but here are some reasons even you, simple reader, can understand that AI regulation is simply a bad idea. We need only look at history.

Let’s start with labor. We can all agree that a twelve-year-old child should not be working in a mine. This has been the law since 1938. The law itself can be traced back to the National Child Labor Committee around the start of the nineteenth century. The members were upset by the millions of children under age sixteen working in factories the prior few decades. In other words, once we put children to work in factories, our society “rapidly” addressed the problem in only a generation or two, just like Sam is suggesting will happen with AI. Ask yourself, where would America be if we had slowed down the industrial revolution by depriving the market of a couple million able-bodied children for the first few decades?

In 1911, after 146 people died in the Triangle Shirtwaist Factory fire, the US sprang into action and improved safety standards in buildings. People had been calling for regulations for years but honestly it felt premature. The universe has its own way of telling us when it’s time, often through mass casualties. When we saw lots of people die, we knew it was time. Sure, premature regulation could have saved those 146 lives (and likely more who died in other fires), but who knows at what cost to business from such premature regulations. A little extra death is sometimes the price of progress.

Next, let’s consider food. Prior to the Pure Food and Drug Act of 1906 food was kept fresh by using formaldehyde, a carcinogen used in embalming. In his speech to Congress on June 21, 1906, Representative James Mann of Illinois said, “Here is a bottle of cherries, originally picked green, in order that they might be firm, with the green color all taken out with acid until they were perfectly white, and then colored with an aniline dye which is poisonous in any quantity.” [emphasis added] How do you like them apples, er, cherries? Such techniques were commonplace back then. There was also plenty of lead, alum, sawdust, arsenic, and plenty of other additives used in the US and Europe throughout the nineteenth century. (source, source) But wasn’t this a small price to pay for the explosion in food options? We went from primarily locally sourced food in the eighteenth century to a national food manufacturing and distribution system in the twentieth all from the food innovations made in the 1800s. Regulations only would have slowed it down. AI promises similar gains, taking us from hand-verified human created and edited stories (read: slow and biased by humans who don’t always follow journalistic standards) to endless AI-generated content (read: more content than anyone can consume, not really fact checked, and “unbiased” because the bias built into AI doesn’t really count as biased according to the people who created it).

Also prior to the Pure Food and Drug Act of 1906 manufacturers were only restricted to make truthful statements about ingredients, i.e., what was in the product; claims as to its effects—good and bad—were not regulated. Honestly, that should have been enough. Critics pointed out that manufacturers could make whatever claims they wanted about the benefits, true or not. You could say snake oil cured cancer as long as it was actual oil from snakes. Nor were safety tests required. Consider the elixir of sulfanilamide; some people got up in arms because over 100 people died from it in 1937. To be fair though, the company did nothing illegal. Sure, the drug contained diethylene glycol, which can be toxic, but it certainly wasn’t illegal to sell untested or harmful drugs. As company founder Dr. Samual Evans Massengill wrote, “My chemists and I deeply regret the fatal results, but there was no error in the manufacture of the product. We have been supplying a legitimate professional demand and not once could have foreseen the unlooked-for results.” (source) Eventually the Federal Food, Drug, and Cosmetic Act was passed in 1938 to address this, so I guess it all worked out in the end, right?

There’s an even bigger impact from lack-of regulations; it was the lack of these acts that spurred innovation. In other words, it's not just that regulations slow down progress, but lack of regulations speed it up by creating demand for innovation. For example, the widespread allowance of low quality, even spoiled meat created demand for Chinese fish sauce to cover up the bad taste. You know this sauce as ketchup, and who doesn’t love ketchup? (The word ketchup is an Americanization of “fish sauce” in the Southern Min dialect.) Early versions of ketchup allowed for things like coal tar; some people don’t like to ingest it but I say if it’s good enough for roofing and paving, it’s good enough to eat. Enter H.J. Heinz, who saw that bad ingredients might not be good and made a better version, with safe, fresh ingredients, including tomatoes, and in doing so built a multi-billion-dollar empire. Heinz ketchup was one of the first products to come in a glass bottle. He did this because he wanted to show the quality of his food product to the prospective customer. Imagine if all food products at the time were safe, he never would have had to innovate. The regulations people are calling for would deprive us of opportunities like this. (Some might point out they would simply create other opportunities, but in a safer world, but we can never know for sure, can we?)

Just in case you’re not convinced, let's consider environmental regulation. The Industrial revolution was really gaining steam in the mid-nineteenth century. The smoke and waste produced led to the peppered moth evolution (the smoke so blackened the countryside that dark moths gained an evolutionary advantage), the Cuyahoga River fires (because of the oil and chemicals that so polluted it), and the melting of the polar ice caps. But we eventually passed regulations like the Clean Air Act in 1963, the Clean Water Act in 1972. Regulations like the Safe Drinking Water Act allow us to turn on our tap water and not worry about what we might be putting in our bodies. All this passed within 100 or so years of the start of widespread pollution.

The point is that if we fix problems too early, we’d add costs and reduce innovation. Sure, many people died in fires and from bad food, and many more got hurt or sick, but all this is in the name of progress! AI has the promise of bringing about changes more rapidly than almost any in human history. And just because changes have been abused in the past, doesn’t mean new ones will be in the future.

I’d like to remind you that there is no direct case of AI killing anyone. Social media has been around for a while and has been shown to cause depression and body image issues in teenage girls. OK, yes eating disorders have one of the highest mortality rates, but hey, nobody said being a teen is easy. And yes, social media has led to widespread bullying. And social media causes depression in adults. But cat videos! And seeing what that girl your liked in high school is doing these days! And vacation photos (the same ones we all didn’t want to see when they were slide shows in the 1980s). No one is saying ban social media, it’s just that some are saying maybe we should have some safety guidelines around bullying and harassment and limiting teen access. But here’s the thing, you have to balance those regulations against the cat videos and the billions of dollars in revenue. A little extra depression seems like a small price to pay for some of the most valuable companies in human history. Had we regulated these sites more, it could have harmed profit. It’s not like social media itself was actively killing anyway.

Bottom line, if you start to regulate too early, you harm growth. To borrow from American Patriot Col. Jessup: These deaths, while tragic, probably helped spur economic growth, and those people who bravely opposed such regulations, while grotesque and incomprehensible to you, created billions, if not trillions of dollars in GDP.

Heck, let’s ask AI! Hey ChatGPT, “Given the historical record of technological changes, is it ok to keep AI unregulated?” [That was my prompt and below was the response.]

‍

The historical record of technological advancements offers valuable lessons about the need for careful regulation. Most groundbreaking technologies—like the steam engine, electricity, nuclear power, or even the internet—eventually required some level of regulation to ensure safety, fairness, and societal benefit. The key argument for regulating artificial intelligence (AI) centers around its unprecedented potential for both positive and negative impacts.

‍

Well, nevermind, it’s still nascent, and nascent technologies never hurt anyone! (Pay no attention to that shoe-fitting fluoroscope behind the curtain.)

Bottom line, you need to ask yourself what is more important: economic opportunities for tech companies and massive profit for their investors, or the health and safety of the general population. Choose wisely, but also choose quickly, because the further down the path we go, the harder it is to change directions.

‍

[Disclaimer: the above is all written in a sarcastic tone, and to quote the above without noting that would be a blatant misrepresentation of my pro-regulation beliefs on AI and other fields. I add this because not only will humans quote me out of context (and those who do should be derided as having no clue what they’re talking about), but AI tools themselves, which DO NOT have permission to scrape and/or train on this content in the first place, but will illegally and blatantly ignore it, also likely won’t get the sarcasm either.]

‍