The image of a malicious, uncaring computer brain wiping out humanity is indeed scary. The machine would be as efficient as it was deadly, exterminating humans even as it coldly ignored their cries for mercy. But if this seems like anything more than science fiction to you, it’s probably because you really don’t understand what artificial intelligence is, or how it works. The above could only ever come to pass if AI was specifically designed to do so by a human. In which case it is the human that is terrifying, not the technology itself.
AI is absolutely steeped in misconceptions, so much so that it might be the most misunderstood technology in human history. Not since cavemen stared in terror at fire have people been so misinformed, and out of touch with how something really works.
AI Is Smarter Than Humans
Intelligence is a very subjective word, and in the case of AI is improperly used. Bingo Australia software is impressive, and indeed capable of offering games of Bingo to players online. But few would think of the software as being intelligent. The software might be intelligently designed, but is hardly intelligent in its operations.
AI works on many of the same principles. It can be designed by intelligent humans, and by extension will act in seemingly intelligent ways. But the software itself is only as smart as its design, and only good at a very specific task. There is no such thing as the mythical Artificial General Intelligence. AGI is the concept of software performing well all tasks, such as the robots seen in movies. AGI does not exist, and likely will never exist.
AI Learns And Acts By Itself
Perhaps the biggest misconception about AI is that it is able to learn, and act of its own accord. Nothing could be further from the truth. AI does not act, behave, or respond at all. It simply carries out a set of predetermined instructions based on programming. The instructions can indeed adjust according to new information, but even this is by specific design, and under very specific, predetermined circumstances.
When referring to ‘learning’ developers are talking about AI systems that can adjust during operation, honing down parameters based on input. This process is automated, predetermined, and in no way reflects decision making or actual learning.
AI Will Become Sentient
There is hardly a more ridiculous and misinformed notion than thinking that artificial intelligence will become sentient. AI software is nothing more than a set of programmed instructions, and although impressive in scope not even remotely comparable to sentient life. Thinking that it could spontaneously become sentient is about as logical as thinking that a book of instructions could wake up and star talking.
AI chat programs, for example, are perceived as intelligent, and are indeed designed this way. But they function as all software does, and are in no way different to other AI. The real issue is that humans bestow intelligence where there is none, then apply the superstition of ignorance.
