Prof Margaret Boden sees AI as a way to understand the human mind, and says the hype is prompting us to think about the human dimension. She spoke to Dr Claire O’Connell.
Suppose you had a robot as a companion, and you were suddenly overcome with sadness or fear. Would that artificially intelligent machine truly know how to comfort you? And should we give robots that kind of responsibility in the first place?
This is the type of scenario that worries Prof Margaret Boden, who spoke in University College Dublin (UCD) this week as part of the UCD Plotting the Future series organised by the UCD Humanities Institute, UCD Institute for Discovery and the UCD Geary Institute for Public Policy.
In particular, she painted the picture of a child being minded by a robot nanny while watching the film Bambi. “What is the robot going to say when Bambi’s mother is shot? The child is going to be distraught,” said Boden, who is research professor of cognitive science at the University of Sussex, where she helped develop the world’s first academic programme in cognitive science.
“In that case, if you knew it was going to watch Bambi, you could cheat, but even if [the robot] says ‘oh it is only a story’, it has no meaning. There is no understanding.”
AI advances
Artificial intelligence (AI) is moving into our everyday lives, noted Boden, and advances such as supercomputers beating humans at chess, Jeopardy! and Go are landmark achievements, enabled by increased computer processing power and data storage.
AI is now also supporting professions such as medicine, law and accounting, and there is speculation about the emergence of artificial general intelligence (AGI) and then artificial super intelligence (ASI) where machines are sufficiently intelligent to improve themselves.
But Boden doubts that AI will reach such dizzying heights any time soon. “I’m not sceptical about computers being creative. They already have been with the likes of AlphaGo coming up with new moves that had eluded humans,” she said.
“But I am sceptical about them having AGI, never mind ASI, within this century. Some people are saying 2030 [for AGI]. Do the maths: that is 13 years away. That is nonsense in my humble opinion. I don’t think people realise how hard the problem is and how rich and subtle human thinking is.”
Hype prompts thoughts
However, there is a silver lining to the hype about AI taking over. It is making us think more about its implications, according to Boden, who is scientific adviser to the Leverhulme Centre for the Future of Intelligence at the University of Cambridge, and to the All-Party Parliamentary Group on AI.
“I cannot think of a single example of human activity that will not be touched in some way by AI, but there are huge dangers too even if you don’t believe there will be superhuman intelligence in 30 years,” she said. “Fortunately, people have started to worry about it, to think about what the problems are and whether they could be regulated. So it is a hugely exciting, problematic and important area.”
Weapons, jobs and human relationships
Some of the more problematic areas she mentions include autonomous weapons that not only find targets but decide to destroy them, and self-driving vehicles that will result in the loss of jobs.
“Many people will say jobs were lost in the first industrial revolution. It was an unpleasant time for some individuals but, for society at large, more jobs and better pay came into existence,” she said.
“[With AI,] new jobs will be created. Who ever heard of a data analyst 10 years ago? But I think those jobs will be open only to relatively highly educated people.”
She also considered the impact on professions where younger practitioners supported by AI could lack the training to deal with intractable problems.
And one of the most worrying aspects for Boden is where computers are put into contexts that should be one-on-one human relationships. That includes caring for children and the elderly, and the ability of experienced doctors to provide human wisdom about personal, quality-of-life issues associated with treatments.
Insight into the mind
Boden has what she described as a “weird” background that saw her study medical sciences, philosophy, and cognitive and social psychology. And, while she learned some computer code along the way, coding is not her thing.
When she wrote one of the first books about AI in the 1970s (Artificial Intelligence and Natural Man), its hundreds of pages did not contain a single line of code, she noted.
“But it explained in very great detail how many different programmes worked, what they could do and what they couldn’t do, and why and what sorts of changes would be needed so they might one day be able to do what they couldn’t do,” she said. “Then there were three chapters on the philosophical, psychological and social implications.”
A fellow of the British Academy and of the Association for the Advancement of Artificial Intelligence, Boden’s abiding interest in AI is what it can tell us about us humans. “AI gives us really good concepts to think about the mind and the brain,” she said. “And the most important lesson that AI has taught us is to appreciate and recognise for the first time the enormous power and subtlety of the human mind.”
Associate professor Patricia Maguire, director of the UCD Institute for Discovery, told SiliconRepublic.com that it was a great honour to host a true pioneer of women in the sciences: “Artificial intelligence is the future and the UCD Institute for Discovery was delighted to host Prof Boden and provide a forum for academics, students and the wider community to discuss an issue that has such enormous implications for all of our futures.”
Prof Margaret Boden’s latest book is AI: Its Nature and Future.
Want stories like this and more direct to your inbox? Sign up for Tech Trends, Silicon Republic’s weekly digest of need-to-know tech news.