There has been a lot of hype about Artificial Intelligence lately, and I think the topic has become out of focus due to misunderstanding what today’s AI really means. To the layman, AI feels like something out of a science fiction movie where autonomous artificial beings are ruling the world. When we read or hear about how AI is here right now with examples being self-driving cars or IBM Watson, it doesn’t connect much with that image in our minds.
AI hype has existed since computers became ubiquitous enough to be recognized as a substantial investment opportunity. Money poured into teams which attempted to make AI on punch cards. Even as technology progressed, the reality always fell short of the hype.
The field of AI research was founded at a conference on the campus of Dartmouth College in the summer of 1956. Those who attended would become the leaders of AI research for decades. Many of them predicted that a machine as intelligent as a human being would exist in no more than a generation and they were given millions of dollars to make this vision come true.
This pattern of dreaming, coupled with funding, coupled with withdrawn funding has repeated many times. When you research the modern state of AI you’ll see mixed results. Here are two Wired articles on the same topic, with different perspectives:
- Obsessing Over AI Is the Wrong Way to Think About the Future
- The Three Breakthroughs That Have Finally Unleashed AI on the World
Artificial Intelligence is actually Intelligence Amplification
The first article is about how today’s AI is actually Intelligence Amplification. What is the difference between AI and IA? Basically, AI is a hyped up version of IA. Today’s state of AI is little more than a tool for amplifying human intelligence. Even the best AI – arguably Watson, is still just a form of IA. You might argue that given Watson’s notorious Chess playing history that it has awareness. That’s simply not the case in the true sense of awareness.
The difference has to do with constraints of problem solving. These advanced algorithms are based on massive amounts of data and parallelization in order to solve problems within specific constraints. Watson played chess well because it was programmed to play chess well. Self-driving cars drive well because they were programmed to apply general AI algorithms to that particular problem.
What is true Artificial Intelligence?
What AI is lacking is awareness. Today’s AI is more of a toolset for helping humans to make better decisions. So what do we make of more recent claims that 30% or 70% or 90% of human jobs can be done by machines? These types of claims have been made for a long time. It doesn’t make them groundless, but AI really can only replace humans within the constraints of particular tasks.
There are reasons AI is behind humans when it comes to the full scope of awareness:
- AI is largely based on the same fundamental way that our human mind works in terms of experience (data) and parallelization (huge numbers of neurons making small decisions in aggregate). This doesn’t mean AI can’t exceed us, it simply means we have been doing this broadly for a lot longer.
- AI does not have the human experience – we import information from our world into the AI world. This means we always have a native advantage – just like humans make poor silicon computers – silicon computers make poor humans.
- There is no inherent motivation for survival in an artificial reality. Natural evolution of life is fundamentally based upon survival. Any feelings of desperation for survival would need to be yet another constraint programmed in.
At the end of the day a computer is a number crunching machine. We can dress it up in seemingly endless and novel ways, but it will still be a glorified calculator. This begs the question of whether humans are glorified biological calculators – that is a topic for another day.
The Future of Artificial Intelligence
I firmly believe that AI with true awareness, intentions, and experience can exist – potentially with our current level of technology. The fact is for these aspects of AI to exist, they would need to be programmed into the software as problems to be solved within constraints – a kind of emulation. What is the value of such emulation to the human species? Perhaps the value is novelty and a kind of shallow companionship.
Personally I can imagine value in the future of AI having true awareness; but, maybe as a species we should be focusing more on world peace and sustainable human existence, before giving birth to Artificial Intelligence akin to humans.
What do you think about the future of AI? Should Artificial Intelligence become more than Intelligence Amplification?