IN a brave new world, some time in the future, life is going to be a dream because computers will do everything, including producing our food and doing all the simple tasks of everyday life that we don’t like doing.
We will live in an era of abundance, with a universal income and driverless cars. We will be free to do the things we like to do, such as hanging out with friends.
That is the optimistic take on the way the world is going, says Roman van der Krogt, an expert in AI (artificial intelligence), who will show some of the applications of AI as they exist today in an accessible lecture at Griffith College on Thursday (March 22), at 6pm, as part of the Cork Lifelong Learning Festival.
Roman, who is Dutch and has done post-doctoral research at UCC, is based in Macroom and works for Cisco which develops networking hardware and telecommunications equipment.
He also presents another, scarier scenario that could occur resulting from the onward march of AI. There is a school of thought “that says the robots will take over” and at some point will become so powerful that we will be like ants to them.
“Why would they want to keep providing for us? Both sides have really convincing arguments. It will be either good in the end or it will end up badly. I haven’t made up my mind. I’m generally an optimistic person but I’m still torn between the two outlooks. I’d probably give more weight to the pessimistic side,” said Krogt.
AI is a complex area. Asked what drew him to it, Roman says: “If a problem isn’t complex, then it isn’t fun. From early on, I was interested in AI and I got a chance to do a doctorate in AI-related fields. Growing up, I would have watched Star Trek, looking at all the computers.”
Roman says there are two types of AI, “narrow or weak AI and general AI”. He explains: “When people think of AI, they typically think of general AI where you have a robot that acts like a human and can do basically everything humans can do. The narrow AI is very good at one specific job.
“Think of your phone, be it an iphone or an android. It understands what you’re telling it to do. Other forms of narrow AI are modern cars that watch for road signs and display them to you and detect people who are overtaking you.
“Driverless cars are already here. if you buy one of the high-end Teslas, it lets you take your hand off the steering wheel and it will do a good bit of driving. It’s very good for highways. The roads in the US are all very wide and easy for driving. But it will take a while before driverless cars will be suitable for the back roads of Ireland.
“Then again, I have a seven-year-old daughter. I don’t think she will need a driver’s license. By the time she can buy a proper new car when she’s 25 or so, she won’t need a license.”
Driverless cars are safer on the roads. But Roman addresses the ethics of self-driving cars: “If a car detects that a collision is unavoidable (say a child suddenly runs on to the road), what is the car to do? Should it steer into the other lane, saving the child but causing an accident between the car and oncoming traffic, hurting passengers? Should it steer in the other direction, onto the sidewalk, hitting the people that are walking there? Or should it do nothing and kill the child?”
Roman adds: “It’s a very interesting thought experiment and, depending on how the question is framed, people will have different, often conflicting responses.
“Germany has introduced guidelines that will have to be followed when it comes to defining future laws on self-driving cars.”
With robots ‘taking over’, what will happen to our jobs? Roman refers to journalism and says that a lot of sports reporting in the US is or will be computer generated.
“The basic blurbs that you read on Yahoo or Associated Press are computer generated. But with all of these things, there are skills. There is always going to be a place for the proper investigative journalist. That is something that needs general AI and we’re still quite far away from it. But the bog standard ‘Man City defeated Chelsea last night by a late goal’ doesn’t require creativity in reporting. It’s standard stuff, as are stock exchange reports.”
Recruitment is an area that uses AI in the US.
“A lot of recruitment companies in America use AI to scan the CVs of people who apply for a job. There might be hundreds of candidates. They can be wheedled down to maybe ten.”
Law, says Roman, is going to use AI a lot in the future: “I have a friend involved in a law company. When there’s big litigation between companies, there’s a discovery phase where people have to produce all the documents that relate to the case. So they let AI loose on a company’s network, allowing it to scan all the documents, emails and network power points related to the case.”
Jobs that are generally “safe” are the ones that have high human interaction.
“So teachers are quite safe, as are social workers. But everything that can be automated, such as factory work, the work of cooks, accountants and even the work of computer programmers, there’s a 50% chance that these jobs will be automated,” he said.
While the work of doctors and nurses requires a big human component, there are already robots that can perform operations. Roman cites an operation on a pig’s intestines that has been carried out in a completely automated way. He says that in Japan, where there is a big elderly population, they are big into designing robots for nursing homes to help people in their daily lives.
“Nothing is really safe, but generally, the more human interaction and compassion needed in a job, the safer it will be.”
Looking at manual labour, Roman says there are robots that can pick tomatoes, for example.
“If you think about it, tomatoes are quite squishy. The robot has to find the ones that are ripe enough. In ten years from now, it will become commonplace for robots to do this kind of work.”
His talk takes place on Thursday March 22, at 6pm in Griffith College.