Dr. Geoffrey Hinton, the “Godfather of AI,” has left Google and joined the growing number of experts sounding the alarm about AI development.
AI development has taken massive leaps in recent months, and tech experts are going increasingly worried about the implications. Some of the industry’s leading names recently signed an open letter calling for a pause on large-scale AI experiments. Meanwhile, some experts have long called AI an existential threat to humanity.
Dr. Hinton has spent the last decade at Google helping the company develop AI, but he’s changing his tune after quitting his job at the search giant. In an interview with The New York Times, Dr. Hinton said he did not want to criticize Google or AI development until he had quit his job, which is one of the reasons he did not sign the above-mentioned open letter.
Now that he has quit his job, Dr. Hinton is mincing no words in expressing his concerns about AI and the potential threat it poses to mankind. In fact, according to the Times, Dr. Hinton says a part of his regrets his entire life’s work.
“I console myself with the normal excuse: If I hadn’t done it, somebody else would have,” he told the outlet.
One of the big concerns with AI is how different individuals, groups, companies, or governments will use it.
“It is hard to see how you can prevent the bad actors from using it for bad things,” Dr. Hinton said.
Another major concern is the AI arms race currently developing between Microsoft and Google. Microsoft — long the runner-up in the search industry — got the jump on Google in AI thanks to its partnership with OpenAI. This led Google to rush out its own AI chatbot, despite many within the company saying it wasn’t ready.
Dr. Hinton says this behavior is much different than how Google acted in the past, when it was a “proper steward,” careful not to release AI products before they were well-tested and safe. Now, however, the two companies’ arms race may lead to something being unleashed that can’t be stopped. This is especially a concern if AIs become smarter than the people that make them.
“The idea that this stuff could actually get smarter than people — a few people believed that,” Dr. Hinton said. “But most people thought it was way off. And I thought it was way off. I thought it was 30 to 50 years or even longer away. Obviously, I no longer think that.”
Ultimately, Dr. Hinton believes scientists should pause or slow development until they develop ways of maintaining control, although he acknowledges such a solution may not be possible.
“I don’t think they should scale this up more until they have understood whether they can control it,” he said.
Dr. Hinton’s entire interview with the Times is well worth a read and should be a sobering caution to users and regulators alike.
