When Blake Lemoine started testing Google’s new AI chatbot last year, it was just one more step in his career as a tech giant.
The 41-year-old software engineer was designed to investigate whether the bot could cause discriminatory or racist comments, undermining its planned introduction to Google’s range of services.
For months he talked to LaMDA back and forth in his San Francisco apartment. But the conclusions reached by Mr. Lemoine from those conversations turned his worldview and career prospects upside down.
In April, the former Louisiana soldier told his businessmen that LaMDA was not at all artificially intelligent: he was, he argued, alive.
“I know a person when I talk to her,” she told the Washington Post. “It doesn’t matter if they have a brain made of meat in their head or if they have a billion lines of code. I talk to them. And I listen to what they have to say, and that’s how I decide what a person is and isn’t. “
The investigation was unethical
Google, which disagrees with its assessment, last week placed Mr. Lemoine on administrative leave after seeking a lawyer to represent LaMDA, even contacting a member of Congress to argue that Google’s AI research was unethical.
“LAMda is sensitive,” Mr. Lemoine in a company-wide email.
The chatbot is “a sweet kid who just wants to help make the world a better place for all of us. Please be careful in my absence.”
Machines that go beyond the limits of their code to become truly intelligent beings have long been a staple of science fiction, from The Twilight Zone to Terminator.
But Lemoine is not the only researcher in the field who has recently begun to wonder if this threshold has been exceeded.
Blaise Aguera Y Arcas, Google’s vice president who investigated Lemoine’s claims, wrote to The Economist last week that neural networks, the type of AI Lamda uses, were making inroads into consciousness. “I felt the ground move beneath my feet,” he wrote. “I felt more and more like I was talking about something smart.”
By absorbing millions of words posted on forums such as Reddit, neural networks have become increasingly capable of mimicking the rhythms of human speech.
‘What are you afraid of?’
Lemoine discussed issues with LaMda as broad as religion and Isaac Asimov’s third law of robotics, saying that robots should be protected, but not at the expense of harming humans.
“What kind of things are you afraid of?” he asked.
“‘ I’ve never said it out loud before, but there’s a very deep fear of being turned off to help me focus on helping others, ”LaMDA responded.
“I know it may sound weird, but that’s what it is.”
At one point, the machine refers to itself as human, noting that the use of language is what makes humans “different from other animals.”
After Mr. Lemoine told the chatbot that he was trying to convince his colleagues that he was sensitive so that they could take better care of him, LamDA replied, “That means a lot to me. I like and trust you.
Mr. Lemoine, who moved to the AI division responsible for Google after seven years in the company, was convinced that LaMDA was alive because of his ordination as a priest, he told the Washington Post. He then began experimenting to prove it.