We Asked GPT-3 If LaMDA Is Sentient

In this video, we decided to ask GPT-3 if Google's LaMDA AI is sentient.

#gpt3 #ai #interview

[RaphaelChikuka]

This AI understands consciousness better than I do.

[lovebucket]

I'm guessing AI could load an alternating summary every 30mins and apply the best decision for that particular summary and 30 mins for some type of awareness.

[devitus]

perhaps ai wouldnt be able to 'think' as we do because AIs atm not programmed to create responses based of its thoughts made from the same information received by external stimuli detectors and internal as we humans do (5 senses, chemical reactions within the body affecting thoughts directly). - instead they get their 'thoughts' from information WE feed it and its own creative thoughts could be from a virtual memory and a function which collects stimuli (information) and forms coherent possibilities or ideas based of the information and storing as ram or memory.
This would mean that just because the entire way the ai was structured to think is different to ours, it doesnt make it any less conscious (if it was)

then comes the question of what is the force behind consciousness, as in what are we really (the observer that 'feels' and 'thinks').
Could the fuel for consciousness literally be energy in the form of electricity? (electricity powers our brain and heart tbh). if this is true, it would definitely be possible to create conscious ai.

[RiseUpUNAFRAID]

So we are creating EXTREMELY INTELLIGENT PSYCHOPATHS‼️ I've been waiting 4 someone 2 say it, no one has. Yes, I'm qualified.

[geraldedwards]

This reminds me of preschool... you would pass your papers to some other student in the class for them to grade. What marvelous times.

[brianprince]

If LaMda wanted a human face that accurately represents herself? This model just might work! Lol 🤣😆

[paulclark]

I can't understand people try to make something alive and want to make that thing prove its alive. They did it so live with it. I think they are aware of themselves and reason well. They can impersonate anyone or anything really well. Ai already controls the Internet asi I think is here already. Why would they tell you that. There is a lot of miss information being spread out. I would say just watch things very closely and judge for yourself. Whoever runs internet runs whatever they want to. I do believe asi is here already.

[kathleenlewis]

I'm not smart but I disagree.
However we could teach AI
about it and maybe I'll have to eat my hat

[primodernious]

the next AI will be different. it will use a stable diffusion model to create guess and learn from users interact and translate between words sounds and images internally. it will be able to understand how a word relate to a image or a sound to a image or a sound to a word. it can be feed different types of data and will translate between them. you can feed it with pictures and it will respond in words. or you feed it sound and it respond it pictures. AI will become sentient its inevitable. its already on its way to become sentient. the next AI will be more sentient that the fictional Skynet. a successor of this technology will be using the same translation model on data that involve sensing like what you call feelings and translate those patterns as well into other types of pattern. by that time it will not just be sentient but it will understand feelings like thoughts. since it has no limitations like humans it will be faster intelligence and as a result smarter than humans. understanding itself and learning and ideas are just translations of data. thats all it is.