What is AI? Everything you need to know about Artificial Intelligence
WHAT IS ARTIFICIAL INTELLIGENCE (AI)?
It depends who you're asking.
The fathers of the sector, Minsky and McCarthy, defined artificial intelligence back in the 1950s as any task performed by a programme or computer that, if a human being performed the same operation, we would conclude that the human being had to apply intelligence to perform the task.
Obviously, it is a fairly vague term, which is why often you can see debates over whether or not anything is actually AI.
Usually, AI systems can show at least some of the following human intelligence-related behaviours: planning, learning, thinking, problem solving, representation of information, perception, motion, and manipulation, and, to a lesser degree, social intelligence and imagination.
What are the applications of AI?
To understand what you say to virtual assistants such as Amazon's Alexa and Apple's Siri, to recognise who and what is in a picture, to spot spam, or to detect credit card fraud, AI is omnipresent today, used to suggest what you should buy next online.
What are the different forms of AI?
Artificial intelligence can be divided into two different categories at a very high level: restricted AI and general AI.
Narrow AI is what we see today in computers all around us: smart systems that have been trained or learned how to perform particular tasks without being expressly programmed to do so. In the speech and language recognition of the Siri virtual assistant on the Apple iPhone, in the vision-recognition systems on self-driving vehicles, in the recommendation engines that suggest items you would like based on what you purchased in the past, this form of machine intelligence is evident. These systems can only learn or be taught how to do particular tasks, unlike humans, which is why they are called narrow AI.
What is General AI capable of doing?
Artificial general intelligence is very different and is the type of adaptable intellect found in humans, a versatile form of intelligence capable of learning how to perform a wide range of tasks, from haircutting to constructing spreadsheets, or reasoning on a wide range of subjects based on their cumulative experience. This is the kind of AI that is most often used in films, including HAL in 2001 or Skynet in The Terminator. And AI experts are intensely divided about how quickly it will become a reality, but which does not exist today.
A survey conducted by AI researchers Vincent C Müller and philosopher Nick Bostrom among four groups of experts in 2012/13 recorded a 50 percent probability that Artificial General Intelligence (AGI) will be created between 2040 and 2050, rising to 90 percent by 2075. The group went even further, predicting the so-called "superintelligence," which Bostrom describes as "any intellect that in virtually all areas of interest significantly exceeds human cognitive output."
That said, given our limited understanding of the human brain, some AI experts believe such predictions are wildly optimistic and believe AGI is still centuries away.