Tuesday, September 15, 2009

Polls Archive 45 - Conscious Computers?

"By 2020 we will see computers that exhibit signs of consciousness that make them able to deal with complex and contradictory input in the same ways that humans do." Poll ended August 6 2009. 35.4% agreed, 9.7% said it was impossible, while 54.9% said it was possible but will take longer than the 2020 target date suggested in this poll.

Wow, a new winner for the most number of votes. The previous winner was Poll 34 - God? Or the multiverse?, which received 258 votes, this one received 277. What is it about the idea of computers becoming conscious that elicits such a strong response in people? If a computer were to achieve consciousness, would that mean that we human beings are not as special, not as uniquely gifted, as we've traditionally believed?

In Logic vs. Intuition, I described how famous physicist Erwin Schrödinger pointed out that within our universe life is a unique process which creates pockets of "negative entropy" or increasing order: and within that context, consciousness and creativity can also be described the same way. In my recent blogs The Quantum Solution to Time's Arrow, and The Statistical Universe, I've also talked about the role the reversal of entropy might play in the creation of our observed universe, and new scientific theories that support that idea. Is consciousness a force that reverses entropy, that acts as a force pushing against disorder?

I would say that if a computer algorithm can become conscious, then the idea that consciousness could exist in other living things, or even within other extra-dimensional shapes and patterns that exist outside of our limited spacetime, all move a step closer towards being accepted concepts. A conscious computer would be a big deal! I find it interesting that so few people responding to this poll thought such a leap will never ever occur.

Last month my good friend. Argentina's Mariana Soffer (of the amazingly diverse blog Sing Your Own Lullaby sent me a link to a blog called "The Mutiverse According to Ben" (a title which of course already makes my ears perk up even before I visit the blog) which introduced me to the term Artificial General Intelligence, or "AGI". According to wikipedia, AGI is also known as "Strong AI", which to my layman's eye looks an awful lot like "Artificial Intuition" as promoted by Monica Anderson, of artificial-intuition.com fame, who I talked about in the original blog that inspired this poll, Logic vs. Intuition. Whether my interpretation is accurate or not, what I see in both approaches is a desire to transcend basic computer logic, and apply creative leaps that better approximate the processes the human mind implements as it deals with the incredibly complex and often contradictory input it receives from the world around it.

Would such a computer be conscious? If a computer some time in the near future said to you "yes, I am conscious", and email conversations with it convinced you that it was conscious, would that mean this ideal had been achieved? That, of course, is the essence of the Turing Test, which I've talked about in entries like 41 - Is Creativity a Quantum Process? , Norway's Reverse Deju Vu, and Computers and Consciousness.

Ultimately, I would say that processes like life, creativity, and intuition are engaged with our spacetime in wider windows that our limited one-planck-length-after-another experience, and that's where the magic starts to happen. Last blog, in Beer and Miracles, I showed you my definition for a miracle: "Any dramatically unlikely occurrence is indistinguishable from a miracle". Next time we're going to look at a poll surrounding another dramatically unlikely event: The Big Bang.

Enjoy the journey,

Rob Bryanton

Next: Polls Archive 46 - Is the Big Bang an Illusion?

Other blogs that talk about unlikely events include:
Randomness and the Missing 96%
Unlikely Events and Timelessness


Skymt said...

I don't think an algorithm can become conscious. It's too simplistic.

But neural networks is another story.

Here's a movie, "In its image", that digs a bit deeper. I found it quite entertaining and interesting, even if I think the speculation is a bit too wild at the end...

Mariana Soffer said...

Thanks a lot for commenting about me in your post. That is nice of you my friend. I was just thinking about what you say regarding consious algorithms. I think that it depends on what you consider that consious is, I think that self-aware algorithms are pretty related to them. There was a famous contest which they used to do every year that consisted in self-aware and self-modifying programs that where prepared to fight-against each other and destroy their oponents.
You can check: http://en.wikipedia.org/wiki/Self-modifying_code

Take care!

Tenth Dimension Vlog playlist