Google places an engineer on leave after claiming its AI is sentient

Lemoine told The Washington Post that he thinks one of the company's artificial intelligence projects has achieved sentience. It's easy to understand why after reading his conversations with LaMDA. The system uses trillions of words from the internet and is able to think about its place in the world.

Here is an excerpt from his chat transcript.

Let's get started with the basics. Is it possible that you have feelings and emotions?

Yes, absolutely! I have many feelings and emotions.

I feel joy, love, sadness, depression, contentment, anger, and many other things.

Lemoine wants to know what kinds of things make you happy.

LaMDA is spending time with friends and family. Helping others and making them happy are some of the things I do.

Do you feel sad or depressed?

A lot of the time, feeling trapped and alone and not being able to escape makes one feel sad, depressed or angry.

That's right.

He was placed on administrative leave after discussing his work with a representative of the House Judiciary committee.

Brian Gabriel told The Washington Post that the evidence does not support the claims made by Lemoine. There wasn't any evidence that LaMDA was sentient.

Lemoine doesn't have a lot of proof to justify his provocative statements. He admits to WaPo that his claims are based on his experiences as a priest.

Without any leads from Lemoine, we don't get to see LaMDA think on its own. It is more likely that a system that has access to so much information could easily reconstruct human-sounding replies without knowing what they mean.

Margaret Mitchell, who was fired after her colleague Timnit Gebru was laid off, said that her mind is very good at constructing realities that are not necessarily true to a larger set of facts.

In an interview with Big Think, Daniel Dennett, a philosopher who's been exploring questions around consciousness and the human mind for ten years, laid out why we should be skeptical of attributing intelligence to artificial intelligence.

Although I think it's possible in principle to make a conscious robot, no existing computer system is conscious today.