Google’s playing games that could have a major impact on your life.
Its AlphaGo artificial intelligence program on Tuesday triumphed over one of the world’s greatest Go players. This was the fifth and final game in a series that began last week, and the win proves in no uncertain terms that the search giant has made great leaps in its ability to predict human behavior.
This is notable because Go is such a complex game — much more so than chess — and because the sort of AI technology used to master the game could impact how people experience life online.
“It’s going to be used for ads. There’s no question,” John Havens, a well-known AI expert and author of Heartificial Intelligence, told The Huffington Post. “I’m not trying to demonize, but for me, the writing is so on the wall.”
The basis of this type of AI is “machine learning.” AlphaGo is designed to “discover new strategies for itself” by playing and analyzing thousands of virtual rounds of Go. But it’s only good at this specific thing — it can’t apply its learnings to a new skill, like playing a video game.
The technology is fascinating — software that can train itself! — but follow the money and you might come to an unsettling conclusion. When Havens says the technology will be used for advertising, he’s referring to the tens of billions of dollars Google makes on ad revenue every year. Google is obviously a tech company, but more precisely, it’s an ad tech company — the majority of its revenue comes from its marketing services.
Machine learning of the sort that informed AlphaGo is definitely applicable to that corner of Google’s business — the company can essentially use software to learn from your personal data in much the same fashion, which increases the efficacy of its marketing efforts by targeting your interests with razor-sharp precision.
So, if you were being cynical, you might say that AlphaGo’s ability to outwit a world-class opponent could translate into Google’s ability to, in a sense, outwit you and get you to interact with its ads. Are there positive uses for the technology? Of course — machine learning could, for example, improve weather predictions and save you tons of grief when traveling. But business is business, some would argue.
“To even think for a nanosecond that they’ve invested the billions of dollars primarily to help — that’s when I get very upset,” Havens told HuffPost.
The problem isn’t so much that you might see more effective ads: It’s that the same sort of technology could shape your life in very meaningful ways. Noted robotics expert Illah Nourbakhsh recently told HuffPost about how political campaigns could leverage similar technology to control voter behavior, for example.
“It leads you to question whether we still have free will,” Nourbakhsh said.
No one’s arguing that you need to break out the tinfoil hats. In reality, Havens is optimistic about the future of AI — but experts commonly suggest that citizens and companies need to define the limits of what society will accept before it’s too late.
For its part, Google formed an ethics board when it acquired DeepMind. Critics, however, complain that it’s secretive. The company did not immediately respond to a request for comment from HuffPost about machine learning and ad targeting.
“We have to trust that what [DeepMind is] doing is the best thing for humanity. And that’s not fair,” Havens said. “Because they’re owned by Google and [its parent company] Alphabet or whatever, one has to assume the big push behind it is to make money.”