Stephen Hawking warned that the development of artificial intelligence (AI) could spell the end of the human race. It could be the biggest opportunity to face mankind, or it could be our biggest threat. One thing is for sure, and thats artificial intelligence will look very different from winning chess matches and quiz shows, as its already been used. It seemed so far off only yesterday, but now we have supercomputers in our pockets, automated military drones, and shopping assistants that greet us with hundreds of facial expressions.
Its thanks to the rise in processing power, and the growing abundance of information and data digitally, we can estimate what AI will be able to accomplish in the future. Today’s research is focused on deep learning which mimics the layers of neurons in human brain, and allows for the crunching of vastly larger amounts of data than current computers can cope with. AI will be a reality once computers are able to perform tasks such as pattern recognition, and translation almost as well as humans will able too. There are already some frighteningly futuristic technology at our hands. Facebook has already unveiled its DeepFace algorithm, which can recognise human faces in images around 97% of the time.
Although this capacity might sound narrow and specific, with todays AI it is accomplished through brute force number crunching. AI programs currently have no way to really understand what gives humans autonomy, and allows them to produce interests and desires. There is still a long way to go before anything suitably accurate to the human brain can be achieved with computer science.
AI is powerful enough today to keep making dramatic differences to human life. Although computers can now play and beat other players at chess, they are still what chess grandmaster Garry Kasparov calls, centaurs. This is an concentrated team of humans and algorithms. This looks to be the type of amalgamation that will be favoured until computers can think for themselves. Supported by AI doctors will be able to use augmented reality, and microscopic data pattern matching to diagnose people who have diseases like cancer earlier and with greater accuracy.
Speech recognition algorithms on ever more powerful smartphones will allow millions of people who are illiterate to access and use the internet from developing countries. Digital assistants will help suggest ideas for hypotheses for academic research in many fields. Image classification algorithms will allow wearable computers such as Google glass to view the world around them and have tourist type information displayed from a detection of certain landmarks.
Although there are positive consequences to be had from AI, consider the less than savoury uses from a democratic perspective. As data is collected by organizations such as the CIA, AI could be used by government to monitor billions of conversations, and emails, and pick out citizens who would say the wrong thing in a world of free speech. This might go so far as to breach the liberties of thousands of citizens worldwide.
AI has broad implications for the workforce in society. Many people will lose out their positions and jobs because of it. Take as an example the original computers. They were often women who would perform calculations for higher up members of organizations in days before a calculator was available. Then the transition to the transistor took place, and we learnt to use calculators, rendering this type of job useless, and many people that the time lost out their livelihood because of it. Training and Education will play a huge role in the migration of white collar workers to new industries and jobs, and the wealth that is created will be geared towards more AI. Soon it will be essential for all workers to have the basic ability to program robots as part of the education.
TV and films have had us believe that the future of AI will be an amalgamation of the Matrix, and Minority Report, with AI invading our lives, and governing us with superhuman cognitive capabilities that conflict with our own. Such artificially intelligent super computers is still a long way off. We still don’t know if its even possible to create them. Despite almost a century of poking around inside the human brain, sociologists, neurologists and even philosophers are still a very long way off from anything more than a basic understanding of how the mind works, and what the definition of one actually is. There is a market for a car that can drive itself to the destination of your choice, but less so for a car that wants to drive to its own destination.
Even if the prospect of and AI envisioned by Hawking is still distant, its prudent and useful for society to plan and how to cope with this new AI. Humans however have been organising complex autonomous entities and systems for centuries. Take for example government bureaucracies, financial markets, and armies. All complex systems that use some sort of autonomy to function. However, we still have to set laws and regulations
This parallel should give comfort to those fearing the future of AI, and they also suggest concrete ways to develop it safely. Ai must be kept open for scrutiny as we already know system designers cannot foresee every set of circumstances that appears, so there has to be a off switch by default. Again though from dangerous devices such as nuclear bombs and traffic signals, humans have designed complicated systems with fail safes. The dawn of AI will be a great one, as long as a little safety along the way is kept to the forefront of the pioneers minds.