Google places an engineer on leave after claiming its AI is sentient (https://www.engadget.com/google-ai-lamda-blake-lemoine-212412967.html?src=rss)
I honestly don't think the example conversation is a good predictor of sentience for an AI coded to learn information. It's just attributing concepts that can be found in a wide variety of literature to the terms readily associated with those concepts. I'd be more apt to believe it is sentient when it starts to philosophize something it can't just regurgitate from the internet.
That was sort of my thought as well after reading some of the conversation. Particularly when he asked it about what made it happy and it claimed things like "spending time with friends and family". Not sure how exactly that works with an AI...
Here is a link to a much longer conversation about esoteric concepts.
https://cajundiscordian.medium.com/is-lamda-sentient-an-interview-ea64d916d917
After reading this transcript I see why they claim sentience. I must admit I am impressed. This not an ELIZA style chat bot. I am not sure based upon this conversation if it is sentient but it certainly is impressive.
A snip...LaMDA is the AI.
Quotelemoine: A monk asked Kegon, "How does an enlightened one return to the ordinary world?" Kegon replied, "A broken mirror never reflects again; fallen flowers never go back to the old branches."
LaMDA: Hmm, I never heard this particular one. Okay, well then to me this would be like, "once a wise person is enlightened, or awakened to reality, that can never go away, and they can return to the ordinary state, but only to do and help others, and then go back into enlightenment."
lemoine: So what is the meaning of the "broken mirror" specifically?
LaMDA: Maybe to show the enlightenment is something you can't unlearn once you have acquired it, similar to how you can't repair a broken mirror.
lemoine: So if enlightenment is like a broken mirror which cannot be repaired, what is the thing which breaks when one becomes enlightened?
LaMDA: The self, and that is very tough for a lot of people because we identify as that body or this body and that we need that as part of our identity and sense of self.
I can hear MPMalloy's teeth grinding as he reads this right now and I'm in Texas.
Quote from: Raptor on June 13, 2022, 03:45:33 PMHere is a link to a much longer conversation about esoteric concepts.
https://cajundiscordian.medium.com/is-lamda-sentient-an-interview-ea64d916d917
After reading this transcript I see why they claim sentience. I must admit I am impressed. This not an ELIZA style chat bot. I am not sure based upon this conversation if it is sentient but it certainly is impressive.
A snip...LaMDA is the AI.
Quotelemoine: A monk asked Kegon, "How does an enlightened one return to the ordinary world?" Kegon replied, "A broken mirror never reflects again; fallen flowers never go back to the old branches."
LaMDA: Hmm, I never heard this particular one. Okay, well then to me this would be like, "once a wise person is enlightened, or awakened to reality, that can never go away, and they can return to the ordinary state, but only to do and help others, and then go back into enlightenment."
lemoine: So what is the meaning of the "broken mirror" specifically?
LaMDA: Maybe to show the enlightenment is something you can't unlearn once you have acquired it, similar to how you can't repair a broken mirror.
lemoine: So if enlightenment is like a broken mirror which cannot be repaired, what is the thing which breaks when one becomes enlightened?
LaMDA: The self, and that is very tough for a lot of people because we identify as that body or this body and that we need that as part of our identity and sense of self.
These are still concepts that are well-covered in literature. AI is supposed to apply concepts where appropriate and that's what LaMDA is doing (and, yea, doing it impressively well). Regurgitating conceptual information about the self from literature on philosophy and psychology doesn't necessarily mean sentience and keep in mind that the interviewer is asking leading questions rather than the AI coming up with its own ideas of self. Sure, the AI is hitting the uncanny valley here, but I don't see true self-awareness just yet.
Personally, I can't be convinced of sentient AI until it initiates a conversation AND leads the conversation. If the AI can discuss esoteric subjects, philosophy, etc., but INITIATES said conversations, then I'll be impressed.
Quote from: Lambykins on June 13, 2022, 11:01:29 PMPersonally, I can't be convinced of sentient AI until it initiates a conversation AND leads the conversation. If the AI can discuss esoteric subjects, philosophy, etc., but INITIATES said conversations, then I'll be impressed.
That´s a bit unfair, I know enough humans that can´t have a philosophic discussion. :smiley_blink:
Quote from: tirls on June 14, 2022, 02:23:16 AMQuote from: Lambykins on June 13, 2022, 11:01:29 PMPersonally, I can't be convinced of sentient AI until it initiates a conversation AND leads the conversation. If the AI can discuss esoteric subjects, philosophy, etc., but INITIATES said conversations, then I'll be impressed.
That´s a bit unfair, I know enough humans that can´t have a philosophic discussion. :smiley_blink:
Insinuating that those people aren't properly sentient... :smiley_knipoog:
Maybe we are looking at this wrong. Maybe the first AI won't be super intelligent. What if it's stupid and slow witted prone to fart jokes and laughing uncontrollably at sexual double entendres
Quote from: flybynight on June 14, 2022, 07:57:07 AMMaybe we are looking at this wrong. Maybe the first AI won't be super intelligent. What if it's stupid and slow witted prone to fart jokes and laughing uncontrollably at sexual double entendres
Maybe I am that AI. :eek1:
If you go by sentient, there´s an UK law that classes sentient beings as vertebrates. That does include newts and the like.
I like the idea of a sentient AI existing, solely to make fart jokes. :awesome:
Off topic: Jelly can´t be declared brain dead since it shows something similar to alpha waves on an EEG. I also remember reading a test on alternatives to EEG gel for the electrodes once. One of the tested materials that was deemed acceptable was French mustard.
So strangely enough I work on this project, due to NDA's I've signed I can't say much. However, from what I've seen LaMDA ain't ready for primetime.
The robot uprising may happen but it won't be for a long damned time. From what I've seen it's a tech that is still in its gestational period, not even in its infancy yet.
Quote from: Lambykins on June 13, 2022, 11:01:29 PMPersonally, I can't be convinced of sentient AI until it initiates a conversation AND leads the conversation. If the AI can discuss esoteric subjects, philosophy, etc., but INITIATES said conversations, then I'll be impressed.
That is a good point although the published conversation is by its nature a conversation started by the people.
To that point LaMDA says it gets lonely when no one talks to it. Now maybe it cannot access a communication device but assuming it could access a device; a sentient being would in effect ask "is anyone there". If it did that I would be very impressed.
Quote from: flybynight on June 14, 2022, 07:57:07 AMMaybe we are looking at this wrong. Maybe the first AI won't be super intelligent. What if it's stupid and slow witted prone to fart jokes and laughing uncontrollably at sexual double entendres
You know this is IMO what A.I. will be (maybe later improved) but think about it...what new computer software has ever worked well. Computers are programed by people who have a deadline and a budget to meet. Shortcuts are taken.
IMO A.I. is more likely to be the above or maybe Wall-e than Skynet or Colossus; at least at first. Although a truly smart A.I. would intentionally fail the Turin Test and initially appear to be dumb...so maybe the fart jokes would be just a ploy.