Following a continuous surge in the number of users since it was first unveiled by its developer OpenAI in November 2022, the AI-powered chatbot ChatGPT faced its first traffic drop last month.
After leveling out throughout May, ChatGPT’s global traffic decreased nearly 10% last month, with the number of unique visitors also dropping, according to preliminary data released by an analytics firm last week.
The numbers are as follows:
- Worldwide desktop and mobile web traffic to OpenAI's website plunged 9.7% between May and June
- Unique visitors to ChatGPT’s website dropped 5.7%, with the amount of time visitors spent on the website down 8.5%
- However, ChatGPT still attracts more visitors than Microsoft’s Bing Chat
How Will This Affect OpenAI?
While the initial hype toward AI tools has seemingly worn off, OpenAI continues to expand its service for paying customers.
ChatGPT’s main revenue comes from subscriptions sold to those who want access to the upgraded GPT-4 model since free users can only use the significantly weaker GPT-3.5 version.
Unlike Microsoft’s Bing, OpenAI isn’t necessarily looking to build a mass-market, ad-supported website.
However, OpenAI provides access to more customizable versions of its software through its application programming interface (API) to other tech companies on its developer website, whose traffic jumped 3.1% over the past month.
Media have suggested various reasons behind the recent decline in ChatGPT traffic.
Some reports suggest the end of the school year could've led to this, with college students on summer break and not using ChatGPT to support their daily writing.
Another reason could be that major tech companies are prohibiting employees from using AI chatbots over risks related to data leaks.
The list of companies banning ChatGPT use for work is long, and it includes giants such as Samsung and Amazon, as well as hundreds of other service providers in the banking and healthcare sectors worldwide.
Why Is ChatGPT So Widely Used by Professionals?
ChatGPT excels in generating conversational responses thanks to its ability to engage in human-like interactions, provide detailed answers, and offer creativity. It can assist with various tasks, from answering questions to generating text.
However, it can sometimes produce incorrect or nonsensical answers, be sensitive to slight changes in input phrasing, and overuse specific phrases.
Earlier this year, OpenAI released an upgrade from its GPT-3.5 to GPT-4.0 model for paid subscribers, which introduced improvements in multiple aspects.
The new model accepts text and image inputs (a feature not available yet), performs at a human level on various benchmarks, and incorporates a code generation option.
It aims to provide more accurate and context-aware responses, but, like its predecessors, it may still generate incorrect or biased information sometimes.