Educators are now faced with a pernicious new challenge spreading through their students’ writing assignments: articles written by artificial intelligence. The first online article generator debuted in 2005. Today, AI-generated texts are found in novels, fake news and real news articles, marketing campaigns, and dozens other written products. The technology is free or inexpensive to use, making it accessible to everyone. And it is already infiltrating classrooms.
The professors at innovate_rye know that this is a freshman biochemistry student, and an “A” student. What teachers don’t know about them is that he uses a powerful AI language model to complete most homework assignments. These would be simple assignments with extended answers. For biology, we would learn about biotechnology and write five good and bad things about biotechnology. I was sending an AI prompt like, “What are the five good and bad things about biotech?” and it was generating a response that got me an A,” says innovate_rye, who asked to use his Reddit handle to avoid being spotted by his university.
Without AI, innovate_rye says homework they consider tedious would take them two hours to complete. Today, this guy’s homework takes him 20 minutes. I like to learn a lot and sometimes, school homework that I have already done before pushes me to procrastinate and not to return the homework. “Being able to do it faster and more efficiently seems like a skill to me,” says innovate_rye.
innovate_rye is not alone. Since OpenAI unveiled the latest application programming interface (API) for its widely used language model, GPT-3, more students have begun feeding written prompts into the OpenAI Playground and similar programs that use deep learning to generate text. The results continue the initial prompt in a natural way and are often indistinguishable from human-written text.
Plagiarism check algorithms do not flag AI-generated text
When AeUsako_ was in his final year of high school last spring, he used OpenAI to generate an entire essay on contemporary world affairs. He explained that although he was unsuccessful in his assignment (he lost points for not citing outside sources), he learned that the plagiarism checking algorithms did not flag text generated by the ‘IA. “Because I used Open AI, I didn’t feel the constant anxiety of having to devote all my time to writing,” said AeUsako_, who also applied to use his pseudonym online.
George Veletsianos, Canada Research Chair in Learning and Innovative Technologies and associate professor at Royal Roads University, explains that this is because text generated by systems such as the OpenAI API is technically a original product generated by a black box algorithm.
The text is not copied from elsewhere, it is produced by a machine, so plagiarism checking software will not be able to detect it and will not be able to pick it up because the text has not been copied from elsewhere. “Without knowing how all these other plagiarism checking tools actually work and how they might be developed in the future, I don’t think AI text can be detectable in this way,” Veletsianos said.
Peter Laffin is a writing teacher and the founder of the Crush the College Essay private tutoring program. He says tools like OpenAI’s are emblematic of other compensation techniques that technology has produced over the past decade, such as cloud-based typing assistants that are supposed to help struggling writers.
In teaching literacy, especially for developing writers, instructors look for the desirable level of difficulty, or the point at which you try so hard not to break but also to improve. Finding the right level, the appropriate level of desirable difficulty of teaching increases their ability to write. So if you’re doing compensation techniques that go beyond finding that desirable level of difficulty and instructing that place, then you’re not helping them grow as a writer,” Laffin said.
Veletsianos notes that it’s likely we’ve passed the point of no return with AI-generated texts, and students aren’t the only ones being seduced.
“We can also start to see where this technology could generate a course on the fly and all kinds of questions around the course. I’m not saying the system we have is the best system, but I’m saying these are conversations we need to and should have to see how we can use these tools to improve not only the effectiveness of teaching, but also its effectiveness and engagement,” he said.
Although Laffin acknowledges that a reassessment of teaching effectiveness is needed, he says it can happen by looking at the types of prompts educators assign to students, noting a difference between regurgitation of facts and discovery of information. However, he fears that products like OpenAI’s text generator make writing a point of detail.
We lose the journey of learning. We may know more things, but we never learned how we got there. We’ve always said the process is the best part and we know it. Satisfaction is the best part. Maybe that’s the thing that got cut out of it all. And I don’t know the kind of person who creates more than anything. Beyond the studies, I don’t know what a person looks like who has never had to struggle to learn. I don’t know the behavioral implications of that, Laffin said.
Training for the future or for the past?
We have recently seen some disturbing information from artificial intelligence. An artist won an art contest with works created by AI, and people cried out about the injustice. Today we have concerns about how students might use AI authors to write papers. They would not be picked up by the plagiarism tools because it would be an original work, but written by the AI. This is a recurring problem with any new technological tool. When it comes out, people complain that the tool, especially for students, should be banned, even though, once the student leaves the educational process, the skill of knowing how to use this tool will be more valuable. than the skill to produce something without it.
In the meantime, innovate_rye is looking forward to the TPG-4, which is expected to be trained on 100 trillion machine learning parameters and could go beyond simple text outputs. In other words, they don’t plan to stop using AI to write term papers anytime soon. I always do my homework on the things I need to learn to be successful, I just use the AI to handle the things I don’t want to do or find uninteresting. If AI can do my homework now, what will the future look like? These questions fascinate me, added innovate_rye.
What do you think ?
See as well :
Medical student caught cheating ‘with Bluetooth device surgically implanted in his ear’ while taking exams
A teacher points to GitHub Copilot as the perfect tool for cheaters during programming assessments, and rekindles fears that the profession of developer will disappear because of AI
An AI writes an essay on why humans should not fear artificial intelligence, in a context where some predict the domination of the machine over the human
How programming students bypass automated submission grading systems to cheat on MOOC assessments