The Three Laws of Robotics need to be overhauled if AI is to power our homes, cars and lives – Wired.co.uk

When it comes to the future of artificial intelligence "only a joint approach will make us strong" says Audi CEO Rupert Stadler, and this involves rewriting the Three Laws of Robotics.

Speaking to journalists ahead of the UN's AI for Good Global Summit in Geneva, Stadler said it's inevitable that artificial intelligence will become integrated into every aspect of our daily lives, but for it to be accepted the public must first trust it.

Audi has already demonstrated its Audi Q7 driverless technology, created with Nvidia, and Stadler plans to demo its all new 'level three' piloted driving with the new Audi A8 this summer. The goal is level five, where no driver is needed. In terms of technological advancement, Stadler believes we are not that far off and predicts prototypes will be demoed from 2020 onwards if public trust is achieved.

Over the past two years, Audi has brought experts in philosophy, psychology, law, and computer science from MIT Media Lab, Oxford University, Singularity University, along with entrepreneurs, together to join its Beyond Initiative. The initiative's aim is to help develop a framework, debate ethical dilemmas that driverless cars necessitate and "make sure AI will share our values when making decisions". One of the most pressing of these problems, Stadler points out, is that we expect technology to always do the 'right' thing, even when it would be virtually impossible for a human to achieve that.

Subscribe to WIRED

Talking about theoretical dilemmas, such as an autonomous vehicle being faced with the choice of harming an elderly person, a pregnant woman, or the passenger in the car, he said: "In such a situation, human beings like you and me have no time for thoughtful decisions. We simply react. Interestingly, we expect the autonomous car to make the right decision, and, quite understandably, people are emotionally touched when thinking of such a scenario. From a rational perspective, such a situation is very unlikely to happen. Of course we as car manufacturers do everything we can to avoid such a situation. Our cars are equipped with many many sensors to detect dangerous situations and to fully brake autonomously if necessary.

"As soon as a car will make decisions by itself in a certain traffic scenario, such a situation can theoretically happen. How should the autonomous car decide when it is not fully clear what will happen in the direction it steers into? Is it ethically sound to choose for the unknown? As a society, we will have to find ways how to deal with these topics. We need an open discourse, in which the massive chances of automated and autonomous driving are considered in relation to the ethical challenges."

To meet these challenges, Stadler suggests we rewrite sci-fi author Isaac Asimov's infamous Three Laws of Robotics. These dictate that AI cannot harm a human, must obey humans unless it means causing harm, and must protect itself as long as that does not involve contradicting the other two laws. Instead, Stadler believes these laws should be: "Number one: we will always handle artificial intelligence based on our human intelligence. Number two: robots and human beings should not harm each other or allow harm by doing nothing. And number three: they should support each other according to their specific capabilities."

Part of Audi's sell for its own future line of driverless cars is providing people with the '25th hour' giving them time they did not have before. Any driverless car should be able to do that. but he suggests that as the technology becomes commonplace, it will more likely see multiple people travelling together in cars for efficiency's sake. Audi will still differentiate itself as the luxury option.

"Maybe you want to take your car alone or do some business. It will be a premium user experience. Maybe it will have an excellent Bang and Olufsen sound system. You could take an hour for relaxation. We are able to hand back to our drivers the 25th hour. Time will be the most precious gift in the future."

Continuing about the future of driving, Stadler said: "The future car I dream of will be a chauffeur who drives me safely wherever I want to go to, a secretary who reminds me what I need to do where and when a butler who gets my groceries, a post box on wheels where couriers can deliver parcels, a private medical staff that keeps an eye on my vital functions and maybe it even becomes an empathetic companion throughout my day. Or in brief: a personal avatar.

"This companion can detect my mood and change the lighting and music and conversation to cheer me up! In a nutshell: AI will allow us to make our lives easier by collecting and interpreting huge amounts of data and by predicting situations of the future. You will be able to play with your children in the car, while the car pays attention to other children playing on the street."

The topic of the summit is AI for good, and, of course, there is plenty of good to be achieved through driverless vehicles beyond the time to relax. Stadler points out that 90 per cent of accidents are caused by human failure, which AI stands to dramatically reduce.

"AI will fundamentally change your mobility, and it's up to all of us to make sure AI is used for the benefit of society. We must set a mechanism for labour markets [to create the] perfect match of man and machine." That relationship between humans and machines needs to be fostered not just in the consumer markets when robot taxis hit the streets and put drivers out of a job, but in Audi's own factories, Stadler said.

"We should not just see the threat, but the opportunities. How the human-machine interface works in a smart factory is always to the benefit of the employee. There's lots of heavy stuff that has to be moved from a to b so why shouldn't technology safeguard employees. And enrichment of jobs will change - there will be different jobs available."

Software engineers and data analysts will be in high demand, and Stadler suggests a basic income could be "the right answer". Humans will still always be better at certain things than machines, such as creativity and empathy, for example.

Yet Audi is definitely not ready to welcome an AI onto its board, as a VC management firm in Hong Kong has already done. "We have the responsibility for 88,000 humans. Sometimes it is good not to be rational alone."

"We have to make sure technology serves society - and not the other way round. Then machines will follow the pace of people again. We want to use AI to secure jobs and to raise the standard of living. At Audi we know: robots wont buy our cars! We have to make sure that our economic system stays in balance. We need a good employment and wealth for our whole economy."

See the rest here:

The Three Laws of Robotics need to be overhauled if AI is to power our homes, cars and lives - Wired.co.uk