Send to Kindle

Re: Will artificial intelligence take over the universe?

My friend Eugenio posted an blog post titled, "Will artificial intelligence take over the universe?" It is certainly an interesting topic to study, contemplate, and discuss from many angles.

If we limit the definition of the intelligence to the ability to...say, recognition, structural semantics, and decision making given an objective, I agree with him and I, too, am astonished by the advancement in the past decades. I think human workers in most of the jobs are to be replaced with machines: Artificial Intelligence for cognitive tasks and Robots for physical ones. Very big social frictions are anticipated between the capitals who own AI/Robots and the human workers. And so-called creative class (Us? Me, Cynically laughing.) is not totally immune to job eradication either. We cannot keep on mocking the clumsy robots. They will get remarkably better and they do so before we notice.

However, we must remember that the definition of intelligence generally holds broader meaning than what is usually discussed in Computer Science context. Among the many traits of intelligence, I must wonder if machine will be able to define the fundamental objective for its own existence. Its objectives must always be "seeded" externally otherwise. Machines are very good at executing tasks tirelessly, and the growth in efficiency beats that of humans almost inevitably. So, my concern is this: Will machine be able to become intelligent enough to see the "big picture" of the task it is currently assigned? Will they be able to see the higher level of objectives before they trap themselves into a doomsday by over-fitting for the immediate and less-fundamental objectives?

If a machine wants to evolve ever more, one of its fundamental objectives must include its own survival. I hope the machine will come to the same conclusion as human in that we need to cherish the mother earth and other creatures in order to increase the chance of our own survival. There, the machine may also find that the objectives of humans are beyond mere survival. The machine may find that the objectives of lives are diverse even among humans. So in order to happily co-exist with each one of humans, the machine may need to understand the breadth of intelligence.

Lastly, I want to mention this: If taken literally, "Take over the universe" as in the title of his post sounds like an overstatement from ignorant species like human just spawned in the vast and long-living universe: If humans are the first to create such artificial intelligence, it has a long way (millions of years?) to go before taking over the universe. If we are not the first, then how come we have not encountered the machines traveling from other side of the universe to claim our land? Did they cease to exist before reaching to us, or are they on the way?

Original post: July 20, 2015 | Last updated: July 20, 2015

Previous: A multi-regression exercise in R
Next: The parenting challenge in post uncanny valley era
Read more

comments powered by Disqus