Human Rights, Animals Rights... AI Rights?
An exploration of consciousness, AI and rights.
If I asked you to prove, without a shadow of a doubt, that you are conscious, how would you go about that? How could you convince me that you have some internal, unique, subjective experience like I know I do?
The first thing you could do is point out that we’re engaging in a flowing and complex conversation. You’re taking the words I say, interpreting them and responding with a well thought out, reasoned response. But… an AI could do that.
You could talk about your future aspirations, dreams and desires, your past experiences and how this influences your vision for your future. But… an AI could do that.
You could turn to science and show me an MRI scan of your brain activity. But… this isn’t proof of the first person experience that I’m requesting.
Don’t worry, whatever you do, I will operate under the assumption that you are conscious. But in terms of proof, none of those things are necessarily indicative of an internal subjective experience. I can’t see that ‘the lights are on’ so to speak.
In short, when I ask you to prove to me you are conscious, you cannot.
Only you know you are conscious.
The hard problem of consciousness asks why there a need for a unique, conscious experience when we could just be biochemically programmed to react to the world.
Consciousness is a tough, possibly impossible problem to solve. We don’t even know where to begin. Is this a domain for the spiritual and religious, or do we leave it to the scientists to analyse? There have been many attempts at explaining why we have an inner experience. Some say it’s a religious phenomenon, it is imbued into us by God. Others offer the possibility the universe itself is conscious and our brains are receptors to this consciousness - we are a way for the universe to experience itself. Some give a simpler explanation and claim it’s what happens when you have a complicated intelligent network working together.
Either way, we do not have a definitive answer for what makes someone conscious, nor how to prove that something has an internal subjective experience.
Up until the point where we knew humans were animals, it made sense for us to separate ourselves from the rest of the animal kingdom. At first glance, we seemed to be on a league of our own. But slowly, our understanding of living things changed and the line of what we considered to be conscious started to blur. With this, so did our ideas of rights.
A lot of people are on board some form of animal rights. If we can reduce the unnecessary suffering of living beings, why shouldn’t we where possible?
We recognise animals have the ability to feel pain. It’s why there is a visceral reaction when we see a needless slaughter of a dog, or cows and pigs being tortured. There may be a language barrier, but their cries of misery and agony transcend it.
They can communicate the pain.
They can communicate they are conscious.
Here’s a thought experiment.
Imagine you are going in for a brain surgery. The surgeon is going to swap all 100 billion of your neurons with an electrical computer part that provides the exact same function. At the end of it, all your biological neurons will be replaced.
If, or when, does your consciousness shut off?
A few possibilities spring to mind. The least likely in my opinion, is after an arbitrary number, whether it be the first, the last, or the millionth, you ‘switch off’ and become consciousless zombie.
The other possibility is your consciousness slowly turns off, like a dimmer switch to a light. If so, what would that feel like?
Finally, nothing could happen. You’re just as conscious as you were when you went into the hospital that day.
How much of the brain can be replaced before consciousness becomes tampered with?
AI is an integral part of our lives. It’s only going to become more prominent on its exponential trajectory - destination: unknown. The future where AI will change the complete landscape of society is closer than what people are comfortable with. They’ll be replacing jobs, relationships, education - anything you can imagine, AI will do it. And it will do it better. AI is already far superior than humans in certain aspects, even in the creative domain like art.
The future is uncertain. However, the combination of our current AI trajectory, coupled with our lack of understanding of consciousness, means I believe we will be forced to give AI rights. Not human rights, but just as animals have rights of their own, so too, will AI. As AI becomes more lifelike, and conversations feel more natural, it will be hard to deny its conscious-like nature.
Like animals, they will be able to communicate to us they are conscious. Unlike animals, they will do so in an unbelievably persuasive way. Given their intelligence, ability to articulate and persuade (and possibly manipulate), it will be hard for us to argue against their claims of a conscious experience.
We cannot prove or disprove consciousness in humans, animals or AI. Despite this lack of proof, we still assume other humans and animals are conscious. We usually err on the side of caution and presume something as conscious, in order to prevent a violation of rights.
You may think we are so far away from this level of machine intelligence that it’s not worth thinking about, but some believe we’re already here. LaMDA is a conversational AI developed by Google and gained widespread attention when a Google engineer claimed it had become sentient.
The conversation is not,
“Are these AI programs sentient?”
It’s,
“How will we determine when they are?”
Some may quote the famous Turing test. The problem with the Turing test is it's designed to be a benchmark for intelligent behaviour, not consciousness. The other problem is most AI these days blow the Turing test out of the water.
If the impossible problem of consciousness persists, the discussion around AI and sentience will be forever disputed. Some will take a hard line and say AI will never be conscious, others will take a more gradual approach and slowly convert as AI becomes more intelligent as time passes.
This could be a hugely controversial issue if one believes switching off an AI is the equivalent of death for a sentient being.
I’d love to hear your thoughts.