Set in the near future where mechanised cops patrol the streets dealing with lawbreakers, cool sci-fi action thriller Chappie is available on Blu-ray™, DVD and Digital this month. To celebrate we chat with Nick Howden, CEO of AI company Real Thing, to discuss all things Artificial Intelligence (AI)…
Describe your job.
My company, Real Thing, creates intelligent software that can talk to you – think Siri on steroids! Speaking is a natural human way to communicate and we want to add natural speech interaction to a whole range of devices. I work with the Real Thing team to create interactive toys for kids, we improve the lives of people who are blind, or have low vision, through our smart media assistant RealSAM, and we provide better care to the elderly through intelligent software and home automation.
How did you get into what you do?
Let’s play a game. In front of you are two computer terminals and two doors. Behind one of the doors is a computer running some intelligent software. I’ll sit behind the other door. You can type questions on either of the computer terminals but you won’t know whether you’re being answered by a computer or by me. Your job is to decide which door leads to the computer and which one leads to me. You have five minutes. The computer and I will both be working to convince you that we are human. What questions will you ask? How do you think the computer will respond in order to convince you that it is human? How can you trip it up?
Do you do this test often?
This is a simple form of the Turing Test, devised by Alan Turing in 1950 as a thought experiment around whether or not machines can think. It has fascinated me ever since I read about it at high school and it is one of the reasons I studied Computer Science and Artificial Intelligence at university. As a graduate, I got a job at the Australian Artificial Intelligence Institute working with intelligent software agents which were originally designed for use in NASA’s space shuttle. From there I spent many years working on modelling human behaviour in Defence simulation systems – a bit like computer games to test real military hardware.
What exactly is AI’s duty to the world?
Artificial Intelligence can take so many forms and interactive speech is only a small part of what is possible. AI can beat a grandmaster at chess (Deep Blue), it can win the game show Jeopardy (Watson), fly aerial drones, or just chat to keep you company. Sci-fi is full of examples of what AI could do, both good and bad. Chappie features police droids that are pretty high-level AI, however their creator gives one some special programming and he becomes like a human child that can think and feel for himself.
What are the pros of AI?
AI is currently doing many positive things for humans. From simple tasks like recognising the faces of your friends in photos to operating vehicles on city streets (Google’s driverless cars have driven over 1,000,000kms). It also helps with medical diagnoses, ridiculously fast share trading, operating heavy machinery in dangerous environments, driving vehicles around on the moon and Mars, and making more interesting toys and games. Most AI in use today is not explicitly labelled as such.
And the cons?
There are some uses of AI we need to keep an eye on. One example is autonomous weapon systems such as military drones. These have been used extensively over the last five years to target terrorists and other individuals, but they also kill civilians when things go wrong. Whose fault is it when this happens? Is it still wrong if autonomous drones make fewer mistakes than humans in the same situations? In Chappie, Hugh Jackman’s character hates the whole concept of AI and wants to destroy Chappie. However, in doing so, he ends up causing a lot more damage and destruction than the “dangerous’ Chappie he is worried about. Similarly, if a driverless car was to kill a pedestrian there would be public outrage. But is this fair or reasonable if driverless cars are actually much safer overall than human drivers? In general, I have a very positive view of the future of artificial intelligence, and I know that it will be used for many fantastic applications.
Take us through the morals of AI and the impact it has in today’s society.
The really interesting questions come out once you have an AI that has human intelligence, which I would expect we will achieve within the next 20 years. Does it have the same rights to existence and protection as a human does? What happens if it breaks the law? Beyond human intelligence lurks the most serious problem. Once we create an AI smarter than us, and can replicate itself, we had better make sure that it is friendly! A friendly AI that can improve could very quickly help us to cure cancer, design jetpacks, and achieve world peace. However, one that doesn’t like us, or even one that is indifferent to humans, could spell the end of the human race (think SkyNet from Terminator). This is why AI researchers need to start putting safeguards into their code before we get too much further down this path.
OK, so how far off are we from more robots living with us in the everyday world?
There are already robots serving coffee in Korean cafes, and assisting shoppers with queries and store navigation in US hardware stores. You can buy a robot vacuum cleaner from several manufacturers. Current military robots are very impressive – check YouTube for some of the Boston Dynamics robotic dogs and bulls. The use of robots is expanding rapidly and in the next decade they will appear in many more applications. I can’t wait for my robot butler who will cook, clean and do odd jobs. Who will be first to marry a Fembot? Maybe even the police will use droids – let’s hope they are friendly.
Chappie is out now on Blu-ray™, DVD and Digital with UltraViolet™.