ChatGPT, the viral conversational AI chatbot created by OpenAI, appears to be losing some of its initial luster and appeal.

After rocketing to immense popularity following its launch late last year, recent data indicates usage and interest in ChatGPT may be declining.

Some longtime users have complained on social media and developer forums that the AI seems to be producing lower-quality responses than just a few weeks ago.

They describe the bot as “lazier,” “dumber,” and prone to more mistakes or nonsensical answers. However, OpenAI denies intentionally downgrading ChatGPT, tweeting that “we make each new version smarter.”

The company speculates users encountering more flaws reflects increased usage uncovering limitations.

Eoin Treacy's view

It would be convenient to lay the blame for falling usage on the fact many students are on holiday. However, the accusation of poorer results is more difficult to explain away. The challenge for large language models is getting beyond the novelty factor.

Providing essay answers is a useful if questionable use case. The challenge is if the short form responses required by most people have to be carefully proofread, wouldn’t be easier to write them oneself? The efforts of media organizations to fire their workforces have been publicly outed as being very error prone. Wouldn’t it be better to encourage employees to deploy AI where possible so they can enhance their productivity where possible?