OpenAI, the creator of ChatGPT has this week offered a subscription service to US users, and also an “imperfect” tool to help detect AI generated text.
ChatGPT is proving to be a popular chatbot and is estimated to have reached 100 million monthly active users in January according to a UBS study, just two months after its launch.
Last month Microsoft confirmed a multi-billion dollar investment into OpenAI to help it utilise AI to transform Microsoft’s tools and services, even as the company cuts thousands of jobs.
But its arrival has already caused problems, with students already caught using ChatGPT to plagiarise schoolwork; long standing issues with racism, sexism, and bias not being resolved; and also it has been found to have accuracy problems.
Cyber security experts are also reportedly worried it could be manipulated into writing malicious software (malware).
However to help teachers and other professionals detect AI-generated text, OpenAI earlier this week launched an “imperfect” tool, which it called “a classifier trained to distinguish between AI-written and human-written text.”
“We’ve trained a classifier to distinguish between text written by a human and text written by AIs from a variety of providers,” it said.
“While it is impossible to reliably detect all AI-written text, we believe good classifiers can inform mitigations for false claims that AI-generated text was written by a human: for example, running automated misinformation campaigns, using AI tools for academic dishonesty, and positioning an AI chatbot as a human,” it said.
OpenAI stressed that its classifier is not fully reliable, and in its evaluations it only correctly identified 26 percent of AI-written text (true positives) as “likely AI-written,” while incorrectly labelling human-written text as AI-written 9 percent of the time (false positives).
The more input text is provided, its reliability apparently improves.
“We’re making this classifier publicly available to get feedback on whether imperfect tools like this one are useful,” OpenAI admitted.
This week OpenAI also launched a pilot subscription plan for ChatGPT for users in the United States.
Called ChatGPT Plus, the new subscription plan will be available for $20 (£16) per month, and subscribers will receive a number of benefits including general access to ChatGPT, even during peak times; faster response times; and priority access to new features and improvements.
It will expand access and support to additional countries and regions soon, but initially it will only be offered to those on a waiting list.
OpenAI confirmed the free version will remain available.
Japanese tech investment firm SoftBank promises to invest $100bn during Trump's second term to create…
Synopsys to work with start-up SiMa.ai on joint offering to help accelerate development of AI…
Start-up Basis raises $34m in Series A funding round for AI-powered accountancy agent to make…
Data analytics and AI start-up Databricks completes huge $10bn round from major venture capitalists as…
Congo files legal complaints against Apple in France, Belgium alleging company 'complicit' in laundering conflict…
European Commission opens formal probe into TikTok after Romanian first-round elections annulled over Russian interference…