Recently, a prominent artificial intelligence researcher declares that machine learning models might be conscious. It caused a huge storm in social media, especially on Twitter. Wanted to express my thoughts in a post here on it.
Before we talk about whether a machine is conscious or not, there are three questions, at least, that need to be answered before we start talking about "Conscious Machines".
First question is, what is the definition of "Consciousness" which philosophers are still debating on it right now. There are many varying definition of it and we need to get it to a suitable concise level that is agreeable to most people first, else all of us will just keep debating whether a machine is conscious or not.
Second question is, assume we have overcome the definition part, is measurement. "How can we measure consciousness and how accurate will it be?" Only if we have measurement, we can determine how conscious a human or machine is. Currently, right now we do have a theoretical measurement called "Integrated Information Index" but measurement is quite challenging right now, with the technology that we have. Perhaps, when technology improves, we can get there.
Last question I feel that needs to be answered is, "Where is the Spring of Consciousness?" How does consciousness comes about in humans and animals? I feel, if we are able to find the Spring, then there is a huge chance of creating conscious machines and if I may even be more controversial and speculative here, we might just stumble upon the Spring of Immortality too!
We have to overcome these three questions first before we can start building Conscious Machines. My guess is if we are able to Define and Measure, we still can build Conscious Machines, but these are just machines providing an illusion of consciousness, a disguise provided by super high computation power rather.
I am keen to think through it more if I have more time but for now but this is what I can think of. What are your thoughts on this?