A chatbot created in 1960 surpassed ChatGPT in the Turing Test in a study conducted by the University of California, San Diego, United States. The experiment attempts to evaluate how well artificial intelligence (AI) imitates a human.
The test, conducted at the university, pitted the ELIZA chatbot, developed by MIT scientist Joseph Weizenbaum in the mid-1960s, against ChatGPT, powered by GPT-3.5, a free version of artificial intelligence.
Essentially, the Turing Test attempts to evaluate a machine’s ability to imitate a human. The analysis does not consider the AI’s “consciousness” but its ability to mimic a person’s communication.
The evaluation was developed by computer scientist Alan Turing in 1950. The test involves three participants: two humans and the machine being evaluated. One of the people assumes the role of interrogator, who is responsible for asking questions to the others, but without knowing who is who.
The University of California evaluation required 652 human participants to evaluate chatbot interactions. ELIZA was able to impersonate humans in 27% of cases, while GPT-3.5 was only able to impersonate humans in 14% of cases.
The evaluators’ decision took into account the machine’s linguistic style and socio-emotional characteristics.
GPT-4 is better than ELIZA
Although ELIZA is superior to GPT-3.5, the chatbot is not as convincing as GPT-4. OpenAI’s paid chatbot was able to fool test participants more often; a success rate of approximately 41%.
Source: Tec Mundo
I am a passionate and hardworking journalist with an eye for detail. I specialize in the field of news reporting, and have been writing for Gadget Onus, a renowned online news site, since 2019. As the author of their Hot News section, I’m proud to be at the forefront of today’s headlines and current affairs.
