OpenAI, the developers of the buzzy conversational AI chatbot ChatGPT, is releasing ChatGPT Plus, a new subscription plan for the chatbot that includes increased access and better performance.
Launched just months after the public preview of ChatGPT was made available, ChatGPT Plus will be available for $20 per month to customers in the U.S., with people from OpenAI’s waitlist getting invited over the next coming weeks. Access and support will then be expanded to additional countries and regions.
The pilot subscription program will include general access to ChatGPT, even during peak times. Due to the chatbot’s popularity, ChatGPT is routinely unavailable due to high traffic on the company’s website.
In addition, ChatGPT Plus will get faster response times and priority access to new features and improvements.
Free users will still be able to use the free version, with revenue from paid users going to support free access availability, the company says.
Since ChatGPT was released as a research preview in November 2022, the company has been receiving feedback and analyzing the chatbot’s interactions to learn more about the technology’s strengths and weaknesses. That has led to several updates, such as improved factuality and mathematical capabilities, better performance, conversation history and other general improvements.
“We plan to refine and expand this offering based on your feedback and needs,” the company says. “We’ll also soon be launching the (ChatGPT API waitlist), and we are actively exploring options for lower-cost plans, business plans, and data packs for more availability.”
In addition to ChatGPT Plus, OpenAI also recently a new classifier to distinguish between text written by a human and text written by AIs from a variety of providers to help detect automated misinformation campaigns, students using AI tools in educational settings and positioning an AI chatbot as a human.
This comes as concerns rise about ChatGPT being used for malicious purposes or for plagiarism.
However, the classifier is only 26% accurate so far, and it incorrectly labels human-written text as AI-written text 9% of the time.
The classifier is publicly available as OpenAI analyze whether “imperfect” tools like this one are useful.
If you enjoyed this article and want to receive more valuable industry content like this, click here to sign up for our digital newsletters!
Leave a Reply