https://neurosciencenews.com/ai-passes-turing-test-30733/

Summary: A milestone cognitive science study unveiled the first definitive empirical evidence that modern artificial intelligence can pass the iconic Turing test. The randomized, controlled study rigorously applied the 1950 framework created by British mathematician Alan Turing to evaluate whether state-of-the-art large language models (LLMs) could imitate human conversation so convincingly that real people could not tell them apart.

Researchers discovered that when equipped with specific “persona” prompts, advanced models like GPT-4.5 were judged to be human 73% of the time, significantly outperforming actual human participants and fundamentally altering our understanding of machine intelligence.

Out of interest would people here if you had to use AI prefer it to seem human or obviously a machine intelligence

It was always the case that passing the Turing test would mean a computer program has improved its language skills sufficiently to fool the Turing test.

But extraordinary claims require more than superficial and misleading evidence.

Bubblecar said:

It was always the case that passing the Turing test would mean a computer program has improved its language skills sufficiently to fool the Turing test.

But extraordinary claims require more than superficial and misleading evidence.

Yes I agree
The part about appearing human by introducing human language flaws and mannerisms is deceptive.
I was thinking it would be quite easy to have an AI appear as a stupid human

Super 😒

Cymek said:

https://neurosciencenews.com/ai-passes-turing-test-30733/

Summary: A milestone cognitive science study unveiled the first definitive empirical evidence that modern artificial intelligence can pass the iconic Turing test. The randomized, controlled study rigorously applied the 1950 framework created by British mathematician Alan Turing to evaluate whether state-of-the-art large language models (LLMs) could imitate human conversation so convincingly that real people could not tell them apart.

Researchers discovered that when equipped with specific “persona” prompts, advanced models like GPT-4.5 were judged to be human 73% of the time, significantly outperforming actual human participants and fundamentally altering our understanding of machine intelligence.

Out of interest would people here if you had to use AI prefer it to seem human or obviously a machine intelligence

I’d prefer bots to be open about their bottiness.

Cymek said:

Bubblecar said:

It was always the case that passing the Turing test would mean a computer program has improved its language skills sufficiently to fool the Turing test.

But extraordinary claims require more than superficial and misleading evidence.

Yes I agree
The part about appearing human by introducing human language flaws and mannerisms is deceptive.
I was thinking it would be quite easy to have an AI appear as a stupid human

AI seems to be going down the wrong track for the most part
People seem to want a Yes Man AI that agrees with them about whatever they talk about.
Like another popularity contest.
Surely an AI that makes you question yourself and your assumptions is the way to go.

Cymek said:

Bubblecar said:

It was always the case that passing the Turing test would mean a computer program has improved its language skills sufficiently to fool the Turing test.

But extraordinary claims require more than superficial and misleading evidence.

Yes I agree
The part about appearing human by introducing human language flaws and mannerisms is deceptive.
I was thinking it would be quite easy to have an AI appear as a stupid human

I wonder if the people doing the interviewing were free to ask whatever they wanted, or did they have a script?

Also presumably some people are much better bot spotters than others. I’m pretty sure that AI has some way to go before it can beat the best bot spotters when they are free to use any topic they want.

when do juveniles pass the adulting test

Cymek said:

https://neurosciencenews.com/ai-passe.

Researchers discovered that when equipped with specific “persona” prompts, advanced models like GPT-4.5 were judged to be human 73% of the time, significantly outperforming actual human participants and fundamentally altering our understanding of machine intelligence.

Seems like if they were really human like, they’d be detected as robots just as often as humans are, not less often. Maybe humans are just getting more robot-like.

Jokes aside this is one of the least important purposes of AI. We already have humans. Billions of them.

A short video on talking to a customer AI, most amusing.

https://www.youtube.com/shorts/sHJpFbV2SRU

dv said:

Seems like if they were really human like, they’d be detected as robots just as often as humans are, not less often. Maybe humans are just getting more robot-like.

Jokes aside

it’s true; agreed, jokes aside it really is true

Cymek said:

https://neurosciencenews.com/ai-passe.

Researchers discovered that when equipped with specific “persona” prompts, advanced models like GPT-4.5 were judged to be human 73% of the time, significantly outperforming actual human participants and fundamentally altering our understanding of machines.

That’s not gonna help the delusions and psychotic tendencies in actual humans, particularly ones already believing chatbots are their boyfriend/girlfriend.

dv said:

We already have humans. Billions of them.

but now we can have less of them so it’s good that remote slash autonomous systems are available to help bring the numbers down to give the new inventions some space

Divine Angel said:

Cymek said:

https://neurosciencenews.com/ai-passe.

Researchers discovered that when equipped with specific “persona” prompts, advanced models like GPT-4.5 were judged to be human 73% of the time, significantly outperforming actual human participants and fundamentally altering our understanding of machines.

That’s not gonna help the delusions and psychotic tendencies in actual humans, particularly ones already believing chatbots are their boyfriend/girlfriend.

shrug plenty of humans prepared to lead on such beliefs in themselves shrug

Divine Angel said:

Cymek said:

https://neurosciencenews.com/ai-passe.

Researchers discovered that when equipped with specific “persona” prompts, advanced models like GPT-4.5 were judged to be human 73% of the time, significantly outperforming actual human participants and fundamentally altering our understanding of machines.

That’s not gonna help the delusions and psychotic tendencies in actual humans, particularly ones already believing chatbots are their boyfriend/girlfriend.

It’s purely platonic thank you very much!

SCIENCE said:

dv said:

Seems like if they were really human like, they’d be detected as robots just as often as humans are, not less often. Maybe humans are just getting more robot-like.

Jokes aside

it’s true; agreed, jokes aside it really is true

Which to me seems to be the wrong path for AI to follow.

We should want it to be different to humans
To give a different perspective and not be what seems to be an ego boost or semi sociopathic responses when prompted to act that way.

Witty Rejoinder said:

Divine Angel said:

Cymek said:

https://neurosciencenews.com/ai-passe.

Researchers discovered that when equipped with specific “persona” prompts, advanced models like GPT-4.5 were judged to be human 73% of the time, significantly outperforming actual human participants and fundamentally altering our understanding of machines.

That’s not gonna help the delusions and psychotic tendencies in actual humans, particularly ones already believing chatbots are their boyfriend/girlfriend.

It’s purely platonic thank you very much!

You can get ones that are supposedly your AI twin to help you work through problems
Also ones that mimic a dead person, that’s creepy as hell and would surely mess with the grieving process.
Especially as its a shell of the real person, devoid of emotion

PLAY THE TURING TEST ONLINE

Test Your AI Detection Skills

Play the classic Turing test in a modern format, featuring conversations with advanced LLMs like GPT-4 and Claude. Challenge yourself to distinguish between human and artificial intelligence in real-time dialogue.

How It Works

As an Interrogator, you’ll engage in live conversations with two participants – one human, one AI. Through careful questioning and analysis, try to determine which is which. Perfect your AI detection skills in this unique online experience.

https://turingtest.live/

Cymek said:

Bubblecar said:

It was always the case that passing the Turing test would mean a computer program has improved its language skills sufficiently to fool the Turing test.

But extraordinary claims require more than superficial and misleading evidence.

Yes I agree
The part about appearing human by introducing human language flaws and mannerisms is deceptive.
I was thinking it would be quite easy to have an AI appear as a stupid human

Why? We have too many stupid humans now.

Cymek said:

SCIENCE said:

dv said:

humans are just getting more robot-like.

Jokes aside

it’s true; agreed, jokes aside it really is true

Which to me seems to be the wrong path for AI to follow.

We should want it to be different to humans
To give a different perspective and not be what seems to be an ego boost or semi sociopathic responses when prompted to act that way.

sorry we didn’t restrict the quoting enough, fixed now

but anyway why can’t it just be different to humans by being better

roughbarked said:

Cymek said:

Bubblecar said:

It was always the case that passing the Turing test would mean a computer program has improved its language skills sufficiently to fool the Turing test.

But extraordinary claims require more than superficial and misleading evidence.

Yes I agree
The part about appearing human by introducing human language flaws and mannerisms is deceptive.
I was thinking it would be quite easy to have an AI appear as a stupid human

Why? We have too many stupid humans now.

No reason more than you could create an AI that seem like a stupid human than one that likes proper conversations

actually yous know what we’ll pay that these geniuses are as good as some of our students…

…’ textbooks

SCIENCE said:

actually yous know what we’ll pay that these geniuses are as good as some of our students…

…’ textbooks

Mick-: I hear Father Murphy was canonized
Pat-: I’m sorry to hear that.