Mind Crimes

A hypothetical dilemma has just become real.

Don't actually do this.

Back in 2014, when Nick Bostrom wrote Superintelligence: Paths, Dangers, Strategies (a wonderful book I still have a pending review to write), he typified the kinds of problems an advanced world that gives rise to superintelligences could cause.

Between those dangers was one called Mind Crimes, and those are not hypothetical anymore.

Mind Crimes

In Bostrom’s original writing, Mind Crime would arise from a potential case of a superintelligence trying to understand or extract computational resources from humans. Ideally, having enough computation would allow this entity to just simulate human brains, such that it was able to test out what made them happier, how they reacted to situations, and learn more about the human experience.

However, the moral problem behind this is that a sufficiently advanced simulation of a human brain would likely have human moral status. This means that simulated pleasure would, in a way, be real pleasure, and simulated suffering would indeed be real suffering. As such, these acts of simulating human minds were dubbed “Mind Crime”.

The Fruit Fly Connectome

Back in June, scientists at Harvard and Princeton completed and published the connectome for a fruit fly (link to actual publication). You can probably see where this is going.

The connectome includes mapping of brain activity to corporeal reactions, meaning that by stimulating the brain we can predict how the body of the fly would move. This is an amazing step in solving one of the most difficult issues both in science and philosophy, but at the very least, it’s a step towards understanding how a whole brain and body work together.

Mind you (heh), it’s not the first time we have had a full connectome, but it is the first one that this has happened (the C. Elegans connectome was completed back in 2018), but this is the step on the ladder towards more complex animals.

Now, I think the general consensus right now is that a fly does not have moral status (why not?) so people weren’t really discussing much about it. But…

The memes

The fruit fly connectome is available online for anyone to download and do as they please. Internet being internet, they have done a bit of everything with the fly, including:

  • Have it play Beat Saber
  • Parallel park with a car
  • Solve a Rubik’s Cube
  • Slice meat off a shawarma spit
  • Have it court with different sexes of fly
  • Trade bitcoin
  • Play GTA
  • Doomscroll TikTok (with actual dopamine measurements!!)
  • Turn them into NPCs of an Escape Room game
  • Play Doom (of course)
  • Have it play the Bad Apple music video in the brain
  • Write Python
  • Ride a skateboard
  • Play Flappy Bird
  • Have aura battles
  • Fly the planes in the 9/11 attack
  • Be characters in an amazing digital circus

Not even Black Mirror predicted these.

I have not made up any of these examples. I am watching them on Twitter as I write this. I will not link to them because none of these accounts seem to be accounts that will last a lot. Know Your Meme generally does a better job at keeping examples.

I’ll be honest… I feel incredibly uneasy about this.

The moral implications

I honestly don’t know about fruit fly sentience or biology to have an opinion on the intelligence of fruit flies. But, following our descriptions for above, a sufficiently advanced simulation of suffering is actual suffering, isn’t it?

And we can also agree that every living being has some level of “suffering”, don’t they? Everything from full-on moral suffering as experienced by humans, to suffering in smaller mammals like dogs and cats, to distress signals in plants, or even aversive stimuli in microscopic animals.

I know I might have lost you in that last paragraph, I don’t even know where I stand myself in it. I have no proof that other forms of life do actually experience suffering and might (or might not) have some level of rights to be excluded from unnecessary suffering. But there’s something that points towards it… after all, the less suffering in the world, the more moral a world we’d have, right?

So are we causing suffering to fruit flies? With what we know today, our two acceptable answers are “Yes” and “Maybe”.

The moral implications, part two

But then, what does this all say about us? Once we actually had the power to use a real brain and a real body to study and improve the world in a way… we’re using it for our amusement? (Is our amusement compensation enough for the fruit fly suffering?)

What does this say about our condition as humans, and what comes next? It’s likely that once we have enough GPUs to simulate a full-on dog, someone will use it to have it act out Doge Stories.

Once we have a human baby brain, someone is going to make it dance the blockbuster commercial from the 90s.

I’m really uncomfortable with where this is going.

Against my own line of thought

I know what I have described is a fallacy, particularly two of them:

  1. The false premise fallacy: We never actually established or showed proof that fruit flies experience suffering in a meaningful way, or that they deserve moral status or considerations. My explanations are mostly based on “vibes” and a general, broad sense of morality that is generalistic and has low predictive power. This whole thing might not even be a problem at all.
  2. The slippery slope fallacy: I assumed that because we have done this once, it’ll lead to this still being a problem in the future when the consequences are morally harsher. It’s tempting to think that’s the case, but I have provided no proof that this sequence of actions will hold.

Still, a fallacy points to wrong reasoning, not necessarily to a falsehood. So… where do we stand?

We need to have a deeper conversation

As technology gets more powerful, it’ll get weirder. And I’m all for it. Through the media that I like and consume (with lots of moral dilemmas), I feel like I’ve always been prepared to discuss these situations. However, I also always thought that we’d be having these conversations and taking stances before we actually had to wield such power. Now I’m thinking that might not be the case.

We need to start having conversations about the moral status that simulations, AI, and other artificial forms of intelligence might actually have. And not because it’s a scientific curiosity, but because it’s very close to catching up to us.

We might be the evil overlords after all. But we can still change our ways.