June 24, 2021

"Artificial intelligence would be very different with more leading women" | Technology

"Artificial intelligence would be very different with more leading women" | Technology



Margaret Boden (London, 1936) began her career at the academy in the early 60s: "A woman there was then a unicorn," he recalls. Today discrimination has nothing to do, he says, but "maybe there will always be more men than women in the artificial intelligence (AI) ", what worries him AI is his specialty in recent decades. Boden is a professor of cognitive science at the University of Sussex (United Kingdom). He has degrees in medicine, philosophy and psychology from Cambridge and Harvard. He does not program algorithms or manufacture robots, but he is concerned about AI, its history, scope and opportunities. He knows the discipline well and speaks with caution about his future because, as he admits, it will be good and bad at the same time, like everything created by humans.

Boden has recently published a short and complex book called Artificial intelligence (Turner) and has visited Madrid in a few days of the discipline of the Fundación Telefónica.

Question. What do you think people think when they hear "artificial intelligence"?

Answer. In a certain type of machine learning, that uses neural networks with multilayers to analyze huge data sets and that is able to notice patterns that were not previously suspected.

P. That within the sector. But when hearing AI ordinary people do not think of intelligent robots?

R. That's bullshit. People have many wrong ideas. There is a sudden fashion with AI. People's ideas are closely linked to fiction and movies. They assume that these things are at least as intelligent as an average adult.

P. This is something that we will not see in the coming decades.

R. Do not.

P. Google Cloud CEO, Diane Greene, he said recently that the AI ​​will be the great revolution of the next 30 years, as the Internet was of the last 30.

R. I would not be surprised There has been an apparent advance in AI in the last five years, although there has been no great theoretical breakthrough. 30 years ago scientists had already found out the bases of the algorithms used in deep knowledge. But they could not develop them for two reasons: one, computers were not fast enough. And two, they did not have enough data. Although I do not think that AI is as important to the daily life of the citizen as the Internet is today.

P. But he said he would not be surprised if someone soon complains to his doctor because he did not consult his AI tool and was wrong to diagnose. That is also affecting the lives of people.

"There has been an apparent breakthrough in AI over the past five years, although there has been no great theoretical breakthrough"

R. This will happen soon, sure before 2030, in the world of health and probably in the financial world. There will be cases where consulting the AI ​​will be a mistake, but also others where it will be not to consult an AI already available. The lawyers will make a lot of money with this.

P. He has said that technology changes us as humans. The AI ​​too?

R. Yes. What I hope is that it does not cause us to be less satisfied with human interaction.

P. The Japanese are already satisfied.

R. They are very different from us. They treat robots differently because they think about the metaphysics of animate beings in relation to inanimate ones in a different way. We make a huge distinction. They do not.

P. He fears, for example, that their use in nursing homes will end up dehumanizing them.

R. My problem is to believe that an AI system can be a genuine partner. That they can have emotionally comforting conversations with seniors who are in the early stages of dementia.

P. He has even asked that this type of investigation not be encouraged.

R. I would not do it. I could do it in ways to use AI to help people in residences. But that particular goal of providing what seems to the person served that is genuine human empathy, would change something that is very fundamental. Someone could tell me: maybe they do not find out, maybe they are so bad that they do not realize it. And this helps them to be more animated, what's wrong? It is not an easy question.

P. Technology will not stop because of human concerns.

R. It is true. There will always be crime, cruelty. But we have to resist as much as possible, educate our children to be aware of it, what else can we do? You will not raise your hands and tell everyone to do what they want. Sometimes we prevent things from being done. It is easy to be very pessimistic.

P. You are not

R. I am aware that huge problems will come. And not all are solvable. I am also aware that huge opportunities will come. So I am very pessimistic and very optimistic. And if they asked me if I had a magic wand with which I could eliminate the AI ​​before it started, I do not know. Would I eliminate nuclear energy with my wand? For the bombs yes. But for the energy? I do not know enough to say it.

P. So far we have learned to live with nuclear bombs.

R. Yes, which does not mean that I would like to live more without them. When my children were small, never, I never told them "when you are older". It was superstition, but I never used that phrase. I was very aware they might not be older.

P. Why could they disappear?

R. My oldest son was born in 1968. The Cold War was still there.

P. Do you say the same to your grandchildren?

R. I have four and I do not envy anyone. They will have a hard life. There will be all kinds of employment problems. They will have to face climate change, water wars, bioengineering and mass migration, which will make the current migration look like a costume party.

P. The number of women in science has increased.

R. Yes, today at least they are not incredible objects, unicorns, which is what we used to be. All my professional life I have faced discrimination. It is something that does not happen anymore, much less.

"My problem is to believe that an AI system can be a genuine partner, that they can have emotionally comforting conversations with older people in the early stages of dementia."

P. But it is not yet comparable to that of men.

R. Maybe there will always be more men than women in AI. First, because young women are still not encouraged to study mathematics. Also maybe, who knows, there could be some biological difference. Men in general are better at facing space and women, with language and emotional relationships. I do not think that's just a matter of education, also of a nature. That may continue to be the case. But it is clear that the AI ​​would be different if there had been more women in the sector. The problems they would have looked at would be different. Today the problems that are looked at are basically scientific, hard mathematical problems, which are not the ones that women would be mainly interested in. They are also interested in more tangled, fuzzy problems related to human beings, which are much more difficult for AI. The AI ​​would be very different with more leading women.

P. Why?

R. The problems would be solved in ways that would respect empathy more. The people who work in the systems of care for the elderly are mostly men. Maybe if it were a majority of women they would look at those problems in a different way. They would be more aware of the issues that concern me. Who knows.

.



Source link