Developer sees human benefits from growth of AI
SAN FRANCISCO — Artificial intelligence will not pose a threat to human jobs — instead, it will make people better, says a developer of an AI robot that defeated some of the best poker players.
Tuomas Sandholm, who is also a professor of computer science at the US Carnegie Mellon University, says a lot of applications involving AI have close links with the lives of ordinary people.
He spoke on the sidelines of the 2018 Connect Conference earlier this month, which is a forum on academic exchanges on AI between Chinese and US enterprises and startups.
Although AI technology, which is widely used in business, cybersecurity and even medicine, is a hot item today, it still faces many challenges — some mature and some less mature, Sandholm notes.
“Clearly, the mature ones are what we see as machine learning, particularly deep learning that is very hot,” he says.
“Strategic reasoning is in a quite different state. It’s much more nascent and you don’t see a lot of applications for that yet,” he says, adding that it is a growth area.
Speaking of the balance between decision-making by AI and by people, Sandholm says humans believe in themselves very much.
He cites the example of Libratus, an AI computer program designed and created by him and his team to play poker.
Libratus was built with more than 15 million core hours of computation and was empowered with an algorithm that computes the strategy by machine learning instead of a fixed built-in strategy of poker games.
Sandholm says when he brought Libratus into a tournament against the world’s top human poker players in January last year, an international team bet that humans would win. And although the human players lost the game in a day or two, the team still believed humans are better than AI.
Sandholm points out that AI is extremely powerful but the technology will not pose a threat to humans in the long run.
“Obviously any technology can be used for good or bad,” he says. “It really depends (on) how you use technology. Clearly, there are military uses. You can imagine that will be bad for mankind.”
However, he refers to his laboratory as a positive example. “In my lab, we make a very big point about making world a better place.”
He says his team is running a national kidney-exchange center for organ the sharing in the United States.
“There was no kidney exchange on the national level until we did it,” he says.
“That saves a lot of lives. In supply chain, our technology has saved over $6 billion in efficiency increases.”
AI technology will not increase unemployment for humans, he says, adding that his kidney-exchange center has not displaced any original jobs and has created many new ones.
“You do more operations and so you need more surgeons and more nurse teams. And more employment has been created from that,” he explains.
When asked about what profession will most likely survive the future, Sandholm jokes: “AI programmers.”
He delivered a keynote address on AI technology and its future application at this year’s Connect Conference, which is held annually in San Francisco’s Silicon Valley.