Human-like behaviour key to AI models passing the Turing Test

The Turing Test, which was developed by Alan Turing in the 1950s to determine whether a computer can display intelligent behavior that is indistinguishable from human behavior, has been passed by OpenAI's GPT-4.5 and Meta's LLaMa models.

This was a significant moment for conversational AI, but it was easily overshadowed by a number of exciting advancements, such as ChatGPT's Ghibli imaging, the search for agentic AI (human-like responses are particularly pertinent for this frontier), advances in AI-based cancer detection, and Google's release of a "thinking" Gemini 2.5 model.

It is one of the most prominent among the more recent candidates, but they are not the first AI models to pass this test. The GPT 4.5, which was released in 2023, demonstrated the most human-like behavior in the testing. It discovered that Meta's LLaMa-3.1-405B (where B stands for billion, defining parameters) and its sibling, the GPT-4o (issued in 2024), were competitors of the large language model (LLM).