He is one of the world’s leading experts on artificial intelligence. Stanford professor, founder of Google X, father of Google’s self-driving car, co-developer of Google Street View, and founder of Udacity, the education platform he is currently focused on. We talked about Sebastian Thrun and during our attendance at the Global AI Summit we had the opportunity to briefly interview him. These are his reflections.
Sebastian Thrun, passionate about the changes that AI brings. Thrun transmits enthusiasm for the advances in artificial intelligence in recent years and smiles when he talks about what is to come, whatever it may be. When we ask him about the state of this technology, he tells us that “it is accelerating. Now we are seeing completely new things that started 30 years ago.”
The first jobs that come to mind are those of his former colleagues at Google. “In recent years we have seen endless advances. From Google Transformer in language to IMAGE, for creating images from text.” He is referring to Google’s alternative to DALL-E or Stable Diffusion. Users haven’t been able to test it yet, but Thrun thinks the same as many: “it’s great that we can have a level of quality like that of real artists.”
Will it improve our lives? It is the great concern of many people. The arrival of such a revolutionary technology presents enormous doubts. Still, Thrun maintains an optimistic message. “It will happen like any other technology. Like electricity, airplanes…artificial intelligence is no exception. Even when AI is used to automate jobs. People will be able to focus on more interesting things. AI will help the people to make decisions and achieve their goals.
Our current problem is in the data. We talk about AI but the focus should be on the data that these algorithms work on. In the end, they are machines drawing conclusions about huge amounts of data. The algorithm can be improved, but experts like Thrun point out that if we want to achieve more precise results with AI, our focus should be on improving the databases. A better database can result in much more effective AI. And it is much easier to achieve than not improving the algorithm or using a more powerful supercomputer.
“I think there is a lot of room in the level of detail of our data. In understanding how we can build better structured databases. That will help us for the future, although there is a lot of work here. There is still a lot of work ahead for the next 20 years,” exposes Thrun.
“10 years is an eternity for AI”, responds Thrun when asked if improvements in this field will continue to grow in the next decade. “Nothing is exponential forever, although right now we are seeing it. There is a lot of material. The machines are going to improve in the next 10 or 15 years and will revolutionize sectors such as medicine or transport.”
Some examples that Thrun describes to us are an AI at the same level as doctors to perform diagnoses or autonomous cars with a level of reliability equivalent to that of drivers. Something that, today, is still far away but Thrun is confident that it will eventually arrive.
Every time he ponders the future, Thrun looks back. “10 years ago we were in the Google X innovation lab. Now Google Brain is a massive, commercial success.” What a few years ago were laboratory ideas and little more than complex mathematical formulas, are now part of the commercial strategy of the big technology companies. AI has gone from being the spearhead of innovation laboratories to being an essential part of any plan for the future.
His greatest fear is just the opposite of that of many. There are multiple Google engineers who have shouted to the sky with AI, convinced that they even have a conscience. However, Sebastian Thrun has a very contrary opinion to those who defend that artificial intelligence has the potential to do things on its own.
“AI is a tool, not something we cannot control. It is not an independent entity that makes decisions freely,” says the professor. But it goes further. Asked what his biggest fear is in relation to AI, Thrun replies that “my biggest fear is that it isn’t understood. That people overestimate what it can do. AI is a very simple thing; it’s finding patterns in data. AI will not become a supervillain, not even in 20 years.”
AI needs to be regulated, but not now. Does this stance mean that AI can thrive freely and unchecked? Not so fast. Thrun is in favor of governments making sure that certain algorithms are not implemented, but he is also an advocate of innovation. “We need regulation, but not now. We have to find a way to avoid accidents, but right now it’s difficult. We have to know how things are going to be in order to know how to regulate them,” explains Thrun, in contrast to the more urgent position of people like Elon Musk , CEO of Tesla and SpaceX.
“My hope is that AI will help us live better lives; prevent us from having to waste time with repetitive processes every day.” A very pragmatic position but that at the same time is combined with the enormous potential of artificial intelligence.