The topic of Artificial Intelligence (AI) has been in the news a lot lately, forecasting doom and the end of humanity. While AI is undoubtedly something very powerful, this feels to me a lot like things are being conveyed in such a way as to convince me to panic. When that happens, it’s time to ask why.
Before we go any further, let’s just agree that there are already a handful of AI companies who stand at or near the top of the pyramid. Anthropic, OpenAI, Meta, Alphabet, etc. There are others, for sure, but these are some of the biggest names in American AI. Representatives from some of these very large companies have been coming out and saying “we need AI regulation, for sure, before this gets out of control.”
On the surface that sounds very responsible of them, doesn’t it? There’s actually a much more self-serving reason these big companies want regulation. Let’s illustrate it with a story.
Picture a Mom ‘n Pop pizza shop. They make great pizza…a neighborhood favorite. Their success is built on great recipes, great service, and a long track record of positive word-of-mouth referrals. After awhile some of the bigger pizza chains start moving into the neighborhood. The Mom n’ Pop pizzeria holds its own for awhile, but the competition is heating up.
Now this is unreasonable, but let’s assume some local politicians decide to regulate pizzerias. They create laws that say “any time you want to sell a pepperoni pizza, you have to make sure each slice has at least six slices of pepperoni on it.” Hmmm. Well, that’s a little intrusive, but okay. So Mom ‘n Pop’s pizzeria hears about this and decides they already do that most of the time, but just to make sure they don’t get into trouble, they start using smaller-diameter pepperoni slices so it’s easier to make sure each slice of pizza naturally gets enough pepperoni.
The politicians, who take campaign contributions from the bigger pizza chains, say “uh uh uh, not so fast.” Then they pass another regulation saying the pepperoni slices have to be at least a certain diameter. Then on top of that, they make another couple regulations governing the sourcing of the ingredients and the length of time you can retain hand-made dough before either using it or discarding it.
It isn’t long before Mom n’ Pop are paying more attention to the regulations than to delivering the quality products that have made them a neighborhood pillar for as long as most people can remember. Come to find out those big pizza chains are making contributions to the local politicians specifically so they can flatten the little guy. It’s the big pizza chains who can afford an entire compliance department solely focused on being familiar with the regulations and integrating them into their business processes. Those chains help write the local regulations. Those regulations become so onerous and burdensome that Mom n’ Pop eventually decide it’s no longer worth the effort. They close down earlier than they otherwise would have because their profit margins shrink, it’s no longer worth the hassle, and the anxiety wrapped up in all of this compliance now eclipses the joy they formerly got from being a neighborhood mainstay. The local pizzerias go out of business and the chains are all that’s left.
In my mind, that’s what’s happening with the topic of AI regulation. In the hands of a skilled user, AI can vastly increase the amount of work a single user can perform. So what if a skilled user in a small start-up company found a way to make AI better in some respect? There are untold numbers of angles this could come from, and it’s very hard to predict. It could help disrupt a trillion-dollar industry. That small start-up could quickly rise to the level of being able to compete with some of the big companies.
That’s not something the big companies want.
So if you’re one of those companies, currently leading (or near the front of) the pack, it’s in your best interest to get some regulations put into place…fast. You do that by hyping the fear factor. You’ll hear a lot more talk about Skynet and Terminator or War Games movies. When there’s enough momentum for regulations, guess which companies get to have input on how the regs are worded? That’s right, the ones at the front of the pack. They write the regulations in such a way that it’s nearly impossible for young start-ups to compete. Only the biggie companies are left with the resources to operate within the regulations and still drive innovation, even if it means the innovation occurs slower than in an unregulated environment.
So do we let companies (and AI) run wild and free? That’s probably not the answer, either. This is an amazing tool; we’re going to need our own advancements in AI in order to stay competitive on the global stage, so if we do impose regulations, they need to be biased toward avoiding the scenario described above, and maybe instead steer development toward what we want and away from what we want to avoid. (We want it to enhance our decision-making abilities and synthesis of information, not do our thinking for us.) We have to assume other countries out there will have fewer regulations and scruples than us, which will allow them to advance more rapidly than we do. How do we compete in that kind of environment? We need to aim for bounded chaos within which the field can naturally evolve.
And obviously, let’s not connect AI to any nuclear weapons. J It should absolutely be used to ingest and synthesize available information during very fluid and high-stress situations (like if we believe we are under nuclear attack), but AI should be used as a decision aid, with its findings quickly and easily able to be verified. It should NOT be given any sort of weapons autonomy, even if it can calculate consequences and courses of action in a fraction of the time any single human can. We have to assume our adversaries are going to use AI to the fullest potential in shortening the decision cycle, so we’re intentionally not competing if we fail to use any and all methods to rapidly synthesize large amounts of information. Let’s keep its role to augmenting our decision capacity, though, not replacing it.
Artificial Intelligence isn’t a bad thing on its own. It IS a game-changer, and we have to consider how we employ it, but I think the fear-mongering is a little much. In the coming weeks, keep an eye on which politicians are pressing for AI regulation. Those are likely the elected officials either preying on constituents’ fears or receiving campaign contributions from large tech companies.