ChatGPT has opened the public’s eyes to both the wonders and horrors of artificial intelligence in the few months it has been available. Used carefully and with good intentions, it is a powerful tool that can stimulate creativity, eliminate blank page syndrome, and speed up productivity. Unfortunately, ChatGPT makes short-term education too easy, as the AI can be used to write documents and take exams. This is where GPTZero comes in.
GPTZero is a ChatGPT detection tool created by Princeton University student Edward Tian, and now you can use it to tell if something you’re reading is AI-generated or not. That is how.
What is GPT Zero?
GPTZero was first announced via Tian’s tweet on January 2, 2023 as a tool to help detect AI plagiarism. Tian noted that the analysis is based on research by a Princeton Ph.D. candidate Sreejan Kumar and the work of the Princeton Natural Language Processing Group.
The analysis is based on some ongoing research. @sreejan_kumar and @princeton_nlp. We hope to publish something empirical soon. but in the meantime this was a fun app to make 🙂
—Edward Tian (@edward_the6) January 3, 2023
Is GPTZero free?
GPTZero is designed for educators, but anyone can use it for free. It is unlikely that this will always be the case. The terms of service mention that it could be Rates for some services. It also mentions a 30-day free trial for new users who sign up.
GPTZero is not an open source project and Tian would have faced high hosting fees if not for the help of Streamlit, a Python application framework with a free cloud service for Streamlit applications. While GPTZero’s traffic is a small fraction of what ChatGPT enjoys, even OpenAI has decided to start charging to recoup its costs.
Although Streamlit hosting is free, it generally limits the bandwidth and resources on this shared platform. This would limit the use of a popular application like GPTZero.
Streamlit is currently helping Tian by increasing the application memory and GPTZero hosting allocation. If it continues to grow at a rapid pace, that could change.
Is GPTZero correct?
There have been claims that GPTZero can be fooled. Following a suggestion from a Reddit user Smellz_Of_Elderberry, I asked ChatGPT to write a short story about the book The old man and the sea like I was a high school student. GPTZero was not wrong.
I tried again, changing the text with some misplaced punctuation, incorrect tenses, and other small errors, but GPTZero still said, “Your text was probably written entirely by an AI.”
The key word here is “probably”, and the developer never promised that GPTZero would be perfect. GPTZero’s accuracy is still being tested, but anecdotally, it seems to work fine for most people.
If you are using GPTZero, it is important to note that errors are possible. When using GPTZero for AI detection or ChatGPT for document writing assistance, you still need to check your work for errors.
How does GPTZero work?
GPTZero analyzes the randomness of the text, known as confusion, and the uniformity of that randomness within the text, which in statistics is called burst. Artificial intelligence is very consistent in its complexity and explosiveness, whereas human writers inadvertently change these characteristics.
The work is not done and Tian notes that more tests will be added to improve the accuracy of AI text detection. In particular, implicit bias is an area that is being explored as another way to detect if a piece of text is AI-generated.
We are still studying the implicit bias in the text generated by LM, so we hope to add more tests and factors to improve the model.
—Edward Tian (@edward_the6) January 3, 2023
How can I use GPTZero?
GPTZero is available on your website or through an application programming interface (API). To use the site, simply copy the text you want to check and paste it into the large tag box Try it.
You can also upload a PDF, Word document, or text file and click have results button. You will also need to check the box indicating that you agree to the terms of service.
Do we really need plagiarism checks?
Beyond the research lab to which many text-generating artificial intelligences have been linked, OpenAI has announced ChatGPT to the public at the end of November 2022. By January 2023, ChatGPT had over 100 million users, making it the fastest growing public app to date.
This means that concerns about plagiarism will only increase as this AI assistance becomes available in all corners of life. Microsoft is incorporating OpenAI technology into Bing search, and Google is testing its own version, known as Bard.
Relatedly, AI imagers like Dall-E and Stable Diffusion are under scrutiny for possible copyright infringement. All of these AI services are trained on the writing, photos and artwork found on the internet created by billions of people.
In a way, artificial intelligence borrows from human intelligence rather than creating it on its own. If I borrow from someone else, I have to give credit and possibly pay a license fee. With generative AI, attribution becomes more difficult as each text or image is broken down into fuzzy elements and then reassembled to create a new piece using thousands or millions of sources.
We either need to rethink how we think about copyright and plagiarism, or find tools to help identify AI-generated material and possibly develop a method of giving credit to the vast number of people who contribute to each AI-generated work.
editor’s recommendations
Categories: GAMING
Source: newstars.edu.vn