Could robots take over? Should we fear a world where robots are smarter than humans? As we moved into the twenty-first century, the world became increasingly digitalised, mirroring fictional visions of the future with robots, instant communication and information sharing. Will the machines we develop work as our slaves or control our governments? A recent research poll of economists, futurists and industrial analysts didn’t produce a clear answer.
Artificial intelligence (AI) is the intelligence demonstrated by machines or software. As scientists have developed software and machines capable of intelligent behaviour, tensions have arisen over how intelligent these beings could actually become — think the movie I-Robot.
So how human could AI become? One argument is based around the concept of experience. As humans, every decision we make at any moment of our lives is governed by our own life experiences. We mentally recall experiences and emotions of any given decision to inform our decision-making. When Thomas Edison was asked about his “failures” in his pursuit of inventing the lightbulb, he responded: “I have not failed. I have just found 10,000 things that did not work.” Whether it’s success or failure, humans are constantly learning. Experimentation is a key component of the advancement of knowledge.
If super-thinking machines are the product of this accumulation of human knowledge, they are essentially people. However, almost all AI has to be provided with common sense by humans. Going back to the idea of experience, artificial intelligence will only be equal to a human’s intelligence when it is able to make decisions based on grounding in real-world experience, learn for itself and extend its own knowledge. This has happened yet, and won’t for decades.
The emergence of artificial intelligence (in a human sense) relies on two main assumptions: that technology in the field of computer science will grow and that there is nothing materially special about the human brain, which allows consciousness to exist. As far as we know, both these assumptions are true, which may mean that machine intelligence is inevitable.