IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Packback Adds AI-Generated Writing Detection Feature

To combat academic dishonesty, an ed-tech company that makes AI-based software tools for moderating discussions and essay feedback is giving them the ability to flag writing that was generated by an AI such as ChatGPT.

ChatGPT creative,Open,Ai,Chatgpt,Chat,Bot,Seen,On,Smartphone,Placed,On
Shutterstock
Amid the growing popularity of AI-based text generators like ChatGPT, the ed-tech company Packback has created a tool for educators that can identify writing that was generated by an AI, potentially helping to combat academic dishonesty.

According to a news release, starting this spring, Packback will embed this feature into its software platforms for essay writing and student-led discussions, called Deep Dives and Packback Questions, respectively. These platforms complement instructor feedback with AI-generated feedback, helping them moderate online discussions and coach students to improve the clarity of their writing. They bear some resemblance to programs like ASSISTments, recently developed at Worcester Polytechnic Institute in Massachusetts to give K-12 students feedback on math assignments.

The news release added that unlike generative AI models like ChatGPT, which generate text on behalf of a user, Packback uses AI to provide students with line-by-line feedback on their own writing, and now it can flag instances where students try to pass off AI-generated content as their own.

"Now that models like ChatGPT exist, it opens up a whole other level of uncertainty. These tools will continue to grow and transform education and the work world, both positively and negatively, much in the same way that the introduction of the Internet and search engines transformed the world a generation ago," Miami Dade College Vice Provost of Academic and Learning Technologies Nestor Pereira said in a public statement. "Now more than ever, it's critical for higher ed institutions to proactively invest in tools that use AI to support instructors. Platforms like Packback are a model for the profound promise of AI in education as a tool to not only help students grow as writers, but also enable instructors to navigate and respond to the impact of tools like ChatGPT."

According to the announcement, the AI detection feature will build upon Packback’s automatic content moderation capabilities which already include plagiarism detection, profanity detection and source-quality detection, which have been major concerns for K-12 and higher ed instructors tasked with teaching a growing number of online students during the pandemic.

It's a problem that has already prompted some schools to take action. Citing "concerns about negative impacts on student learning," the New York City Department of Education banned ChatGPT from city school-owned networks and devices last month.

"As models like ChatGPT only continue to grow in popularity, detection of AI-generated writing is only one part of the picture. Higher education institutions, K-12 schools and the education technology community alike have an opportunity — and a responsibility — to design new assignment types that coexist alongside these emerging models," Packback CTO Dr. Craig Booth said in a public statement. "Tomorrow's world of work will be one where professionals collaborate alongside generative AI, which means it's more important than ever to help students build the skills of critical questioning, editing, curation and self-reflection. We're excited to continue exploring the role that instructional AI can play in helping today's students navigate a rapidly changing and increasingly digital world."