Artificial Intelligence. How long until Skynet?
The prevalence of smart phones and tablets means we are increasingly dependent on the connected world.
Part of that connected world is the voice recognition software that runs on these devices: Siri on Apple, Google Voice on Google devices and the new Alexa on Amazon’s Echo device.
These virtual assistants are useful, but do they represent a real artificial intelligence? The simple answer is, “No”; they interpret a spoken voice command and give an answer based on a web search. In a way, they create the illusion of intelligence.
Real artificial intelligence (AI) is best defined as intelligence displayed by a machine. To do this, the machine must be seen to perceive its environment and then take decisions and actions that give it the best chance of succeeding at some task or goal. In effect the machine is able to mimic the cognitive functions that we associate with the human brain, such as problem solving or learning. The key disciplines that current AI research is focussing on are:
- Language processing
- Perception (using sensors to monitor the surrounding environment).
In the past five years or so the quest for true artificial intelligence has significantly advanced. A good example is the emergence of the self-driving car which makes all the decisions needed to successfully drive on a busy road. In theory these cars will drive much more safely than cars driven by humans. In fact, 13 of the 14 accidents the test cars have encountered over 1 million miles driven have been caused by human error.
One of the reasons that so much progress has been made in recent years is the massive increase in processing power of modern computers. Very advanced algorithms can be processed in real time meaning decisions and solutions are arrived at almost instantly.
Artificial Intelligence – The Last Accomplishment of the Human Race?
One of the key features of artificial intelligence is that options to achieve a specific task or goal are analysed logically and without any emotional interference. This lack of emotion should mean that the most efficient and effective solution is reached every time. However this approach does come with some controversy; emotion is considered by some as a key ingredient in the human decision making processes. For this reason experts feel that empowering machines with artificial intelligence could have serious consequences for the human race.
Films like ‘Terminator’ and ‘The Matrix’ have popularised this theory with harrowing looks into a future controlled by machines. Only recently the renowned physicist Stephen Hawking suggested that the creation of advanced AI could be the last accomplishment of the human race if we do not learn to understand the risks associated with it. Other well-known exponents of a cautionary approach to AI are Bill Gates and Elon Musk.
The main threat comes from the possibility that bestowing software with learning capability could mean it exponentially improves itself to the point of surpassing human ability. This “technological singularity” (the hypothesis developed by Vernor Vinge) means that future events and impact on human civilisation are completely unpredictable.
Regardless of the moral implications, the development of artificial intelligence is marching forward at a relentless pace. Big tech organisations like Intel are investing billions of pounds into the technology. This will have huge implications on humanity as we move forward. Not only will we have self-driving cars, but many other systems we take for granted will have intelligence. This potential to make decisions could make roles undertaken by humans obsolete – and that’s another debate.
As with much technological innovation, one of the key sectors aiming to exploit artificial intelligence is the military. Drones that can seek out targets independently will soon be controlling the skies above the battlefield.