Roboroaches and Sentient Simulations
“Backyard Brains” are set to begin shipping their newest version of the “The RoboRoach“, the “world’s first commercially available cyborg”. The package includes a “backpack”, a battery, some electrodes and a cockroach to stick them in.
The suggested “surgery“ involves putting the cockroach in some ice water for a few minutes, gluing three electrodes onto it’s head, sticking a needle into it’s back, inserting a wire into this hole and finally cutting both of the cockroach’s antennae and sticking two more wires inside them. Then wait a while and your RoboRoach is ready to be controlled from your smartphone.
This article was one of many to raise ethical concerns: “When their antennae are trimmed and electrodes are stuck in, they lose that natural sensing ability and it is replaced with an electric impulse sent from a phone. There is little explanation on the Kickstarter page of any lasting effects this surgery has on the cockroaches.”
Images of chopped-up, super-glued cockroaches (and worms) quickly come to mind. But $99.99 is a lot to pay, and people seem to think cockroaches are unpleasant, so hopefully it won’t sell very well.
But what if you had virtual cockroaches?
Well, not exactly virtual cockroaches, but Grand seems assured that his creations will have at least some form of consciousness and will be much more sophisticated than one of his previous creations, Norns.
“When they look at you, there will hopefully be someone behind their eyes in a way that wasn’t true for norns”, says Grand.
In the comments section of one of his blog posts Grand also remarks on the Grandroid’s potential consciousness: “The new brain model is MUCH closer to a mammal’s brain. It’s extremely complex but assuming I can make it all work, the creatures will actually THINK, in the sense that they attempt to predict the future, they can combine sequences of actions to assemble a plan that they think will lead to a desired outcome, simulate possible narratives in their minds, have hopes, fears, worries, intentions, expectations, dreams, and so on”
Grand seems reasonably concerned with their welfare, and writes that he hopes “to worry a little more about Grandroids” than he did about norns.
However, Grandroids introduces itself on it’s Kickstater by saying that:
“Imagine that someone has discovered alien creatures on a nearby planet and is planning to import them to Earth to sell as pets. Would you want to buy one?”
The idea is worrying: create some conscious artificial organisms and then let people play with them without any rules whatsoever?
How should we act if sentient simulations are created? Allow anyone (which might include sadists, animal abusers, etc) to create as many conscious organisms as they like, and impose no restrictions on what they do to them?
Obviously this would be insane, but nevertheless it seems like a plausible outcome.
Hopefully Grandroids wont really be sentient. Given the lack of success in creating sentient computer programs so far, we can be hopeful that Grandroids wont be conscious. But looking to the future, we certainly seem on a track towards artificial sentience at some point.
Another project which moves us closer to artificial sentience is OpenWorm. OpenWorm aims to create a virtual C. elegans nematode worm. Luckily C.elegans has a very small nervous system, which gives us reason to hope that it’s also not sentient.
But I doubt this will be the last endeavor of this type, it seems likely that attempts to simulate organisms with larger and larger nervous systems will follow.
It looks like OpenWorm might also be played with, but there is another way sentient simulations might suffer greatly: by being used in ecological simulations.
Life in the wild is incredibly gory, creating simulations to study population dynamics or evolution would cause huge amounts of suffering to the organisms in the simulation. Many simulations of evolutionary processes already exist so it seems likely that sentient simulations would also be used in this way.
Hopefully neither Grandroids or OpenWorm will be sentient, but there’s no doubt that the potential for sentient programs to be used unethically is huge. You probably won’t be successful enough to sell a million cockroaches and electrodes to stick in them, but it’s much easier to simulate a million virtual ones, and maybe some simulated predators would be next.
This post originally appeared on Ruairí’s blog here: http://www.ruairidonnelly.com/?p=230