- When Chatgpt uses the GPT-4.5 model, it can pass the Turning Test by fooling most people into thinking it’s human
- Nearly three -quarters of people in a study believed that AI was human during a five -minute conversation
- Chatgpt is not conscious or self -conscious even if it raises questions about how to define intelligence
Artificial intelligence sounds pretty human to many people, but usually you can tell quite quickly when you engage in an AI model. However, that may change as Openai’s new GPT-4.5 model passed the Turning Test by fooling people into believing it was a human during a five-minute conversation. Not just a few people, but 73% of those attending a University of California, San Diego Study.
In fact, GPT-4.5 surpassed some of the actual human participants who were accused of being AI in the blind test. The fact that AI made such a good impression of a human being that it seemed more human than actual people say a lot about the shine of the machine or just how awkward people can be.
The participants sat down in two back-to-back conversations with a human and a chatbot, not knowing which one was, and had to identify AI afterwards. To help GPT-4.5 succeed, the model had a detailed personality to emulate in a series of requests. It was asked to act as a young, slightly awkward, but internet expert introvert with a variety of dry humor.
With the little pushing towards humanity, GPT-4.5 was surprisingly convincing. As soon as the requests were removed and AI went back to an empty slate personality and history, the illusion collapsed. Suddenly, GPT-4.5 could only fool 36% of those surveyed. The sudden nosedive tells us something critical: this is not a mind that wakes up. It is a language model that plays a role. And when it forgets its character sheet, it’s just another AutoComplete.
The clever is not awareness
The result is undoubtedly historic. Alan Turing’s suggestion that a machine capable of talking well enough to be confused with a human being may have human intelligence have been discussed since he introduced it in 1950. Philosophers and engineers have struggled with the Turing test and its implications, but suddenly the theory is much more real.
Turing did not respond to passing his test with proof of consciousness or self -awareness. That’s not what the Turing test really measures. The fact that the nail vibes from the human conversation are huge, and the way GPT-4.5 induced actually human interaction is impressive, right down to how it offered mildly embarrassing anecdotes. But if you believe that intelligence should include actual self-reflection and emotional connections, you are probably not concerned about the AI-Infiltration of humanity yet.
GPT-4.5 doesn’t feel nervous until it speaks. It doesn’t care if it fooled you. The model is not proud to pass the test as it does not even know what a test is. It “knows” only things that a dictionary knows the definition of words. The model is simply a black box with probabilities wrapped in a cozy linguistic sweater that makes you feel comfortable.
The researchers made the same point that GPT-4.5 was not conscious. It occurs, not perceived. But notions that we all know can be powerful. We cry on film. We fall in love with fictional characters. If a chatbot delivers a convincing enough action, our brains are more than happy to fill the rest. No wonder 25% of Gen Z now thinks that AI is already self -conscious.
Of course, there is a place for debate about this. If a machine speaks like a person does it matter if it is not one? And regardless of the deeper philosophical implications, an AI that can fool that many people could be a threat in unethical hands. What happens when the smooth speaking customer support rep is not a Harried trainee in Tulsa, but an AI trained to sound disarmamented useful, specifically to people like you, so you pay for a subscription upgrade?
Maybe the best way to think is for now as a dog in a suit that goes on the hind legs. Of course, it may look like a little entrepreneur on the way to the office, but only human training and perception gives this impression. It is not a natural look or behavior, and does not mean that banks hand out business loans to canines soon. The trick is impressive, but it’s still just a trick.
