Artificial Intelligence: What Is AI?
What is Artificial Intelligence?
Artificial Intelligence is basically what its name implies, it is a form of intelligence artificially created by humans. Usually, a thinking machine, a computer, is created to perform a specific task, like to calculate equations, file your taxes, or give you directions to the nearest sex toy store in your city that's open after 11pm. Technically, a computer doing any of the above tasks is exhibiting a limited kind of "artificial intelligence" if they are in fact performing these tasks correctly. In many cases, a computer or computer program designed for one specific thinking task will perform that task much better than even the best humans, that's why computers are so widely used.
But when we talk about the philosophy of AIs, we're imagining a theoretical computer that does more than figuring out your taxes or telling you how to get to Maggie's World of Colossal Dildos. We're talking about the idea that a thinking machine could theoretically possibly become so human-like, we would one day not be able to tell the difference between communications from a human or from the machine. That's what most people are talking about when they talk about artificial intelligence, or AI.
A Brief History of AI Research
In 1950, Alan Turing, a pioneer scientist in the field of computer science, posited his famous Turing test, that basically said that a machine might be able to be considered thinking if it can carry on a conversation that would make the human it was having a conversation with unable to distinguish whether they were talking to a machine or a human.
In the 50's research also began on programs that are called Game AIs, or programs designed to outperform humans at specific games like chess and checkers.
The 1960's saw a lot of research into AIs, including trying to make programs that could do algebraic word problems, analogy problems used in IQ tests, and have dialogs with humans (natural language processing).
However, then there was what was known as AI Winter. This was a dying off of interest in and funding for AI research. A lot of it had to do with encountering difficulties and setbacks, such as the problem with computers translating language without the machine being able to understand nuances of meaning. In 1973 in the UK, an influential report saying that AI research could not meet its lofty objectives led to a reduction in its funding. A lot of the reduction of AI interest in the 70's proved to be hype backlash; AI scientists had either made grandiose claims about what their products could do, or the media took what they said out of context, creating a public expectation that the computers rarely met.
While some people think of AI winter as the end of AI research, Wikipedia notes that:
"Technologies developed by AI researchers have achieved commercial success in a number of domains, such as machine translation, data mining, industrial robotics, logistics, speech recognition, banking software, medical diagnosis and Google's search engine." - AI Winter
Now, most AIs people hear about in the media are "chatbot" style AIs. These take things people say to them and try to put it together in comprehensible language, so that the chatter talking to the AI cannot tell if it's a human or not. However, these are currently limited. They don't seem to possess the human general intelligence that goes into speaking, listening, and processing linguistic information by relating it to memory.
All the chatbot can do is select a bit of information that it thinks is connected somehow with whatever the person just said. It doesn't even seem to be able to stay consistent, meaning that you could ask it the same question multiple times and get multiple answers. Or you could have a conversation and hear the chatbot AI directly contradict something it said a few minutes ago! I'd say we're a million miles away from having a conversation with a bot that closely resembles a conversation with a human.
Are AIs depicted accurately in fiction? Usually, if an AI is showing up in fiction, it's as the villain. AIs can lull humans into a false sense of security by feigning benevolence, and then withdraw such benevolence when the humans decide to do something the too-powerful AI does not like. This is usually done to serve as a moral to the story, about the dangers of hubris, laziness, or trusting too much in the machine. Some fictional AIs include:
- Hal 9000, from 2001: A Space Oddyssey, probably the example most people think of when they think of AIs.
- The ship's autopilot system, or "auto", from Wall-E, based on Hal 9000.
- The Magi system in Neon Genesis Evangelion. It's interesting because it is wired in with the brain of its original programmer, Naoko Akagi, the mother of Ritsuko Akagi, the current chief scientist at Nerv. Therefore, the Magi is like an AI but also is fuzed with an actual human brain.
- The Tachikomas, robots originally used as weapons and military transport in Ghost in the Shell, become self-aware and have philosophical conversations with each other in the show, Stand Alone Complex.
- Jane in Speaker for the Dead by Orson Scott Card and its sequels.
- In Battlestar Galactica, the Cylons were once ordinary machines that became sentient.
- In the Matrix films, the titular Matrix happens because robots once made to do menial labor develop a human-like level of consciousness and begin to struggle against humans for their political rights. It doesn't end well for humans (see the Animatrix).
- A few show up in Star Trek from time to time.
- GLaDOS in Portal.
So in reality, we can't even make a computer that you could chat with for ten minutes without realizing it's a computer and not a person. The capabilities of AIs in real life are severely limited, and their handicaps will take decades of research to solve. However, this does not stop fiction writers from saying, "What if?", and speculating about what might happen if a computer program or robot could achieve sentience similar to that of a human.
Do you hope to see generally intelligent, human-like AIs in the future? Or do you think we've gone too far in trying to create a machine that can think like a human? Let me know by commenting! Thanks for reading!
Wikipedia: Artificial intelligence.
History of artificial intelligence.
Philosophy of artificial intelligence.