OpenAI Launches ChatGPT, Sparking a Global Conversation About AI

OpenAI released ChatGPT, a free research preview of a conversational AI built on its GPT-3.5 language model, letting anyone chat with the system through a simple web interface.

Within five days, OpenAI said more than a million people had signed up to try it, as users shared examples of the model writing essays, debugging code, drafting emails, and explaining complex topics in plain language. Unlike earlier GPT-3 demos aimed mainly at developers, ChatGPT's conversational format and free access made large language models tangible to a mainstream audience for the first time.

The launch also renewed scrutiny of AI's limitations and risks. OpenAI acknowledged the model could "sometimes write plausible-sounding but incorrect or nonsensical answers," a behavior that quickly became known as "hallucination." Educators raised concerns about student use for assignments, while investors and rival labs began racing to respond, setting the stage for a wave of generative AI competition throughout the following year.