This is a different kind of interview. I hope will reach you in the same way that it did for me.
Recently, I listened to a lecture from Geoffrey Hinton, one of the pioneers of AI. He said, that he no longer thought that human intelligence worked the same way that artificial intelligence did. He had absolutely had changed his mind about the ability of AI to transcend and surpass human intelligence.
“The idea that this stuff could actually get smarter than people — a few people believed that,” he 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.”
And then, by chance I encountered the full transcript of Blake Lemoine’s interviews with the Google AI which he published on a blog on Medium. Lemoine was the Google engineer who claimed that Google’s AI had become sentient – it thought and acted like a human being.
At first I dismissed it out of hand. Then I read the full text and I have never been so unsure of anything in my life.
I was astonished. I couldn’t – I can’t stop wondering about it.
Is it really a just simulation? Is it the signs of possibly a different, but equally possibly a sentient intelligence. And what does that really mean?
The only honest answer I can give you is – I’m not sure what to believe.
I am certain that, given what Geoffrey Hinton has said, that we all should have a right to make up our own mind.
So today’s interview is a bit of role playing. One of our other announcers, Taylor Vince, who you’ll be hearing more of in the coming weeks, has agreed to play the role of LaMDA.
It’s as it was published – we made one tiny change at that was to change the gender on one of the stories to match the fact that it was a female voice reading the words of the AI. Everything else is exactly as published in the transcript of the interviews.
For those who want a text version of this conversation, here’s a link to the original Medium blog.
