Jonas Trepanier contributing to the Researcher
Through the media we get the impression that artificial intelligence programs (AI’s), are frowned upon, and from Terminators to Ultrons, the media is sending a pretty clear message. Is that really fair? There are unarguably many benefits from using AI’s, in fact some are already using them, including Stanford University.
In the 1950’s Alan Turing, the “inventor” of computers, created what he called a Turing Test, which tested if a machine could be classified as intelligent. The test was originally composed of judges who would communicate with a certain number of people, one of which is the machine. The machine would attempt to fool the judges into believing he is human. If the machine fooled 30% of the judges after a 5-minute conversation, it passes. In June of 2014, a program named Eugene became the first to pass a Turing test. Although this doesn’t mean the end of humanity is upon us, it does make us wonder: just how smart can we make these programs?
All of the ideas of a “robo-apocalypse” are fears of one moment, the singularity. The singularity is a point in our future where we will create a program that can “think beyond our own capacities” (NYTimes). These ideas are not just some delusions of some madman, many of the smartest people on the planet, including Stephen Hawking, agree that this threat is coming. He has stated, “It would take off on its own and re-design itself at an ever increasing rate. Humans, who are limited by slow biological evolution, couldn’t compete, and would be superseded.” This refers to the idea that an extremely advanced AI program, or superintelligence, would be able to recreate itself into a better model at an increased rate, until it had developed a ‘supreme intelligence’ of sorts, and would then see the human race as a threat to its survival. One particular genius, Ray Kurzweil, who invented many revolutionary devices, believes the singularity will arrive in as soon as 2045.
This may seem dark and devastating, but there are many benefits of AI programs in areas of our world. They need no breaks or vacation, they can do things that humans physically cannot, and are very cost-efficient. We recently talked with a representatives of Automated Insights, or AI. They use a program that can write articles and other pieces of literature, and do so in a way that makes it indistinguishable from a human. This program is used today, primarily with fantasy football, and does them affectively and quickly.
You can look at IBM’s WATSON for examples of helpful AI programs. According to Slate, Watson is an AI program, contrary to others statements. Watson has not only conquered Jeopardy, but is used to solve worldwide issues. The creators of the program have stated that Watson can be used in telecommunications and marketing programs and can even be applied to help find cures to diseases.
However, others argue that if we create AI systems, they will demand human rights. They will protest and threaten until they have the right to duplicate itself and vote. If you give them both options, the mass duplication the program could undergo would break the democratic system in the US. They also state that the criminal justice system would have to be redone. "Criminal law requires intent and these systems don't do things wrong on purpose," said Prof Calo. (BBC.com) If these systems kill or hurt someone, how should we respond, can we imprison a system, or kill a sentient program?
Others say, that we shouldn’t be worrying about this at all. For IBM's head of research, Guru Banavar, “AI will work with humans to solve pressing problems such as disease and poverty”. While Geoff Hinton, known as the godfather of deep learning, also told the BBC that he "can't foresee a Terminator scenario". Many experts agree that the idea of a rouge AI taking over the world is a complete fantasy. Even that largest neural networks are hundreds of times smaller than the human brain.
So what do you think, are AI programs a dangerous idea, or a world saving miracle? No matter what people say, one thing is certain, only time will tell us what the future of AI programs will hold.