ChatGPT
Former special assistant to President GW Bush Pippa Malmgren reacts to artificial intelligence technology and the Russian-Ukrainian conflict on ‘Making Money’.
OpenAI, the San Francisco-based research company behind ChatGPT, said it has released a new tool to help distinguish between AI-written and human-written text.
The company announced an initial version of the build on Tuesday, saying it aimed to gather feedback and share improved methods in the future.
The creators of ChatGPT warned that it is impossible to reliably detect all texts written by AI. However, the company believes that good classifiers can help flag automated misinformation campaigns, position an AI chatbot as a human, and use AI tools for academic dishonesty.
The launch of “Al Text Classifier” comes after a week of discussions in schools and colleges over fears that ChatGPT’s ability to write just about anything on command could fuel academic dishonesty and hinder learning.
POTENTIAL GOOGLE KILLER COULD CHANGE THE US WORKFORCE AS WE KNOW IT
Teenagers and students were among the millions of people who started experimenting with ChatGPT after it launched on November 30 as a free app on the OpenAI website. And while many have found ways to use it creatively and safely, the ease with which it can answer take-home test questions and help with other homework has caused panic among some educators.
As schools opened for the new year, New York, Los Angeles, and other major public school districts began blocking its use in classrooms and on school devices.
A ChatGPT logo seen on a cellphone. (CFOTO/Future Publishing via Getty Images/Getty Images)
The longer a passage of text, the more effective the tool is at detecting whether an AI or a human has written something. Type in any text – a college admissions essay or a literary analysis of Ralph Ellison’s “Invisible Man” – and the tool will label it as “highly improbable, improbable, uncertain if it is, possibly or probable” generated by the AI.
CLICK HERE TO GET THE FOX BUSINESS APP
But just like ChatGPT itself, which has been trained on a huge amount of digitized books, newspapers and online writings, but often confidently spits out lies or nonsense, it’s not easy to interpret the result.
The Associated Press contributed to this report.
