Let’s get down to basics first. What is artificial intelligence?
To understand “artificial intelligence” we will have to start by defining what “intelligence” is. In science we have a definition, that we use when we try to talk about different animals, when we try to understand which ones are more intelligent. Intelligence is the ability to do the right thing at the right time, to recognize an opportunity or a crisis in an environment and to then do something about it. So by that definition even plants are a little intelligent. Because they notice things, such as one side getting sunshine and the other not getting sunshine. So they grow towards the sunshine and may even drop leaves on the other side. Lots of people find that an unsatisfying definition of intelligence, which is where we can start having an interesting conversation.
When does it start to be artificial then?
It is not that intelligence suddenly becomes artificial. The word “artificial” means that we, human beings, have decided to build it. Otherwise there is no difference. A thermostat is also intelligent in the way that a plant is intelligent. The difference is only that we built the thermostat for a purpose.
Is a refrigerator which beeps when the door is left open for several seconds an example of artificial intelligence then?
By that kind of definition that I have described, yes. So what you have there is a sensor which is built into the device to observe its environment. It is a very simple form of intelligence, like a plant’s intelligence. It is not cognitive and not learning from you. Now, if somebody puts a CPU – a little computer that does the thinking — into your refrigerator, then your cooler will turn into the Internet of things that could gather data about you.
Is this the part where the refrigerator tells me that I am about to run out of eggs? Learning from what I put in and take out, right?
Again it depends on how you do it. So you could have a refrigerator that could just detect the eggs, then there is no memory except for the refrigerator itself. But the other way to do it is to keep a record of how often you buy eggs. And so it starts recording information about you and it knows that this is the day you usually buy eggs and put them the fridge and wonders hey, where are the eggs? This fridge now is, what I was just saying, not just intelligent but also cognitive. It has memories about you; it can make predictions about you. This is where someone can get information from your refrigerator whether you bought eggs this week or not and whether you are home or not. That is the point where you will have to worry about your cybersecurity. And this is why it is important to have that simple understanding of AI, to realize that a refrigerator or any Internet Of Things’ equipment could possibly give out information about you. An ordinary refrigerator, which is not connected to the internet, has no security problems.
Some experts say more than 80 percent of Americans, and I think many in other countries as well, are using AI unknowingly. Can you at least tell me what these things are?
Everyone who is holding a power steering wheel is holding AI. Similarly, anyone who has a smartphone of course has AI. And again, it could be simple AI that never improves, it could be AI that is learning the words you use and predicts the names in your address book. Web searches, spelling check, grammar checking, writing programs, that’s all AI. AI isn’t only a robot with a gun. Using your mobile phone to search what weather you have is using AI and also you have to know that other people might also know what you are searching for.
So, should we be worried?
We should be excited and worried at the same time. Mostly AI is making us more powerful. Because of AI there is a reduction of a global poverty. It is incredible; 4 billion people have gone out of extreme poverty in the last 15 years. And a lot of that is through mobile phones. So we have to say, this is a wonderful opportunity. But also there is this terrible and scary part of it. Look at the recent event in China. Millions of people are detained in re-education camps [in Xinjiang] and are not allowed to communicate outside their borders. Two years ago President Xi Jinping made the world believe that China is going in a positive direction. And that’s the danger of AI, it empowers all of us, but someone with a lot of power can use it and turn it into autocracy. So I think we have a lot to work on, doing interviews like this and in policy meetings.
What are the DOs and DON’Ts when it comes to AI?
It is not the AI that you need to worry about. The thing you have to worry about is what memory is being stored and how this memory is protected. So if something is not connected to the internet then there is little that can go wrong. But that includes telephone contacts stored on the Internet. If you have a laptop … always update your security. Once you are on the Internet, you need to keep your security high. If you have a cheap phone and you can not upgrade it then make sure you are careful about the information you put on the phone. Just realize that that information can be taken by someone. Be aware where your cameras and microphones are. One very important thing that people should be worried about is that AI can be used also to fake real-time videos. And this is coming.
Back to you now, when was the first time you got in touch with AI?
The very first time? That is a good question. Well, there are two things I remember: when I was a kid in the 1980s, I applied for a physics camp and I was among the chosen to go on a Saturday morning. And when we reached it they [the organizers] gave us their computers. On the computer was a computer game called Advent, which is short for adventure, but back then the names could only be 6 characters long. So that was my very first experience playing computer games, and computer games are AI. It wasn’t a normal machine where you have to follow their scripts. This thing would make a move and respond to what you were doing and it was incredible. But probably my first exposure was also watching Star Trek with my dad. When my mom went away with her girlfriends, my dad and I would stay at home and watch Star Trek.
Who was the first scientist that you admired the most?
The first scientist I knew of was Jane Goodall. I was amazed by her work. She was the person who made me realize I could be a scientist. Before that I guess it was when I was 4 or 5 when I learned about dinosaurs at school. I came back home and told my dad, who is a regular engineer, that I wanted to be a scientist who studies dinosaurs, and I asked him what that is called. He said, “A paleontologist.” And so for years when people asked me what I wanted to be I would say, “A Paleontologist.”
How many women are working in this area?
What I noticed while doing my PhD at MIT, which is a good engineering school, was that there were more women in AI than in computer science but still not a lot. However, if you look at certain sub-fields like developmental robotics, I don’t know why, 60 percent of them are women. When I was trying to understand intelligence, one of the things I did was working with primates. I think in primatology, actually, there are probably more women, and it isn’t easy to notice them. At the very top of the field there are more men. I find it interesting that nowadays more women working in this sector are making themselves visible. That’s cool.
Tags: AI, Education, Innovation, Insights, Research, Science, Skills, STEM, UK