Since Microsoft-backed OpenAI released “ChatGPT” (An advanced AI chatbot) earlier this year, people across many industries, from engineers, to artists, and writers, are now concerned that their job is at risk due to how quickly ChatGPT can generate incredibly detailed, human-seeming conversation, answers, and even full essays. Soon, it became clear to most educators that the primary use-case for ChatGPT was, unfortunately, as a tool to write academic essays and cheat in online exams.
The power of ChatGPT is staggering, and offers immense value for our society, but it cannot come at the cost of our own human education. Entire essays can be drafted and submitted in a matter of minutes, even on books and topics that a student has never read, complete with citations and quotes in any format. ChatGPT is also somewhat well-informed, with knowledge of many classic books and cultural phenomena. While there are clear limits to how realistic and accurate ChatGPT’s responses can be, the fact is that just a few minutes of editing from a student can completely resolve these errors. Furthermore, the rapid speed of technological advancement means that it is nearly impossible to distinguish an essay written by ChatGPT and modified by a student, from an essay written by a student. Several tools popped up in response to this wave of ChatGPT plagiarism, claiming to be able to distinguish ChatGPT-written content from human content. While these were partially effective with the first generation of ChatGPT, within weeks, OpenAI released the next generation of ChatGPT, called “GPT-4”, which is extremely difficult to detect. What’s more, just a few simple edits by students can easily bypass any of these detection tools. Given the pace of advancement, it seems that this kind of detection technology will never be able to distinguish with the certainty required for an academic judgement.
Data published by OpenAI, illustrating the difference between the sophistication and performance of different generations of ChatGPT, demonstrates just how quickly the technology has advanced in only a matter of months. It’s also worth noting that one of the key metrics on which OpenAI is testing their technology, is how easily it can pass a variety of typical educational exams, such as the AP tests (on which ChatGPT scored nearly all 5’s), and more.
The typical model of plagiarism detection, such as that employed by well-known and widely used tools like Turnitin.com, stores and processes student content to determine if it is similar to other work or has plagiarized elements. Because ChatGPT is effectively now undetectable, and because all the content it creates is original, this leaves instructors in the classroom with few options with tools using this typical model now rendered completely ineffective. Many instructors are choosing to adapt by trying to create ChatGPT-proof assignments, asking about specific things in a text that they believe ChatGPT won’t know, or asking students to do shorter, in-class writing assignments. Other schools are looking for ways to completely block ChatGPT on their network, or on student devices, which excludes students from accessing ChatGPT for other purposes, such as exploration, software development, fun, and sparking new ideas.
Authoriginal has developed a simple solution to this problem, that not only completely prevents the use of ChatGPT and other AI assistants on writing assignments, but also prevents all other forms of plagiarism as well – including the use of ‘purchased essays’ which has long plagued instructors. Here’s how it works – students simply download a Google Chrome extension that creates an Authoriginal Document in Google Drive. Whenever students are actively writing on this document, Authoriginal detects whether any flagged sites, such as AI assistants, are being used. The Authoriginal Document also prevents copying and pasting from other sources (excluding citations), and prevents access unless the student has the Chrome extension enabled. Although Authoriginal is simple, it completely prevents the use of ChatGPT, other AI assistants, and any other form of plagiarism as well. This means that with just a simple Chrome extension, instructors can verify that every students’ work, is really their own. Authoriginal is also non-intrusive and student friendly, enabling students to write their essays anytime they want, with detection only happening when they’re actually writing. All this is LMS integrated, meaning that instructors just create a submission link in Canvas that only accepts Authoriginal Documents, and students submit their essays just as easily as before. Using Authoriginal is a simple experience, but it’s a game changer for education.
We're releasing a free tier of Authoriginal that can help instructors everywhere immediately implement a solution to ensure that ChatGPT is not disrupting education, while preserving access to all of the useful and important things about this new technology. Regular and school-wide use of Authoriginal will likely benefit from the Authoriginal Pro plan, which enables unlimited author submissions, 24/7 support, and screenshots of flagged activity which are critical for making an academic judgement.
Please reach out to us at sales@authoriginal.com, or call us at (818) 641-5936 to chat about getting started with Authoriginal Pro. Authoriginal’s free tier is not live yet and is currently being released to a small number of schools on a pre-release basis. Make sure to sign up for news and updates on our site if you want to be the first to know about the free tier, or send us an email or call! We look forward to hearing from you!
Get a Demo of Authoriginal
