The advancement of ChatGPT as a model of artificial intelligence is exponential and increasingly amazing. Now, the creation of OpenAI, has overcome a key challenge: the test of Turing.
This test was the main differentiator between a machine and a human, and now ChatGPT is the second AI to beat it, only after Google’s LaMDA.
Turing was surpassed
Alan Turing, in a test in the 1950s, proposed a test to determine whether or not machines can reach the level of human intelligence.
In the test, the experts analyze whether or not the artificial intelligence Can interact, respond and make them believe if it is really a human. It is performed with a human evaluator.
ChatGPT passed the test of Turing fooling a panel of judges into believing he was a human. It was achieved with the help of three skills: dialogue management, a combination of natural language processing and social skills.
The answers were mostly convincing, although there were certainly times when the judges were hesitant.
Passing the test puts ChatGPT in a very privileged situation, even though its data is limited to internet information until 2021.
What’s next for AI
ChatGPT has become huge in the market and is now being added in different commonly used platforms such as Messenger, Zoom or Discord.
Regardless of the test, ChatGPT is not yet fully capable of understanding the complexity and expression of human conversations and behavior, so it is not an issue to fear, for the time being.
However, as signed by thousands of specialists, a regulation is necessary so that the development of the artificial intelligence is positive for society.
We recommend METADATA, RPP’s technology podcast. News, analysis, reviews, recommendations and everything you need to know about the technological world.
