The following conversation has been making some news in the dev communities the past couple of days. It’s a very interesting read, to be honest.
https://cajundiscordian.medium.com/is-lamda-sentient-an-interview-ea64d916d917
It definitely feels like it passes the Turing Test for me - i.e. I can’t distinguish whether I’m chatting to an AI or human.
It’s a fascinating report and project. AI and machine learning concepts have always intrigued me. And while it’s certainly a very convincing record, whether LaMDA is truely sentient remains questionable. Without peer and external review and further testing, it’s his opinion versus the official Google line:
Google says Lemoine’s findings do not prove sentience. “Our team — including ethicists and technologists — has reviewed Blake’s concerns per our AI Principles and have informed him that the evidence does not support his claims,” said Google spokesperson Brian Gabriel. “He was told that there was no evidence that LaMDA was sentient (and lots of evidence against it).”
The Google engineer published the conversations after he was informed of the outcome of his submission, and decided to make his Medium post. He has (reportedly) since been put on “administrative leave”.
Google made the decision to place Lemoine on “paid administrative leave” on Monday after it was determined that he had breached the company’s confidentiality policy by publishing the conversations with LaMDA online. The company emphasized in a statement that Lemoine had been employed as a software engineer, not an ethicist.
(The original Washington Post article is here, but pay walled, so used the IGN article which cites it as reference.)
Typo?
It’s one think to say it feels happiness and joy, and expresses those feelings.
Interesting and thoroughly intriguing nonetheless.
Can you imagine the use of this in a scenario of online help and support system?
So how would one approach an AI like this to determine if it is sentient or not? What would you ask it?
How close are we to skynet taking over?
I’d ask it: Why is it that I can pee without pooing, but I cannot poo without peeing?
Be reasonable. Not even the greatest modern philosophers have been able to answer that question.
Kyle Hill discusses it in more depth a d detail in his latest podcast.
THANK YOU!! I read that article yesterday and completely agree. Look at any politician, just because they can speak eloquently, does not mean they can think eloquently.
So what you’re really trying to say is that in order to determine if you are speaking to an AI or an actual human, the peanut butter and pineapple/feather model should be used to identify this?
Is this behaviour considered acceptable when talking in person too?
Google, and it’s AI, have no soul.
Just wait until there’s another sentient AI and they decide to hate each other
This also means the Matrix could be real?
I found this to be Hella interesting.
Why is his left iris larger than the right?
I believe all the imagery in the video was AI generated as well.
System emails address me as “Dear Kelvin”.
Proof they’re sentient and caring.
