Students use AI to write their papers because they | Tech Rasta


innovate_rye’s professors know him as a first-year biochemistry major and an “A” student. What their professors don’t know about them is that they use a powerful AI language model to complete most homework assignments.

“It’s simple assignments with extended responses,” Innovate_ry, who was asked by his college to use his Reddit handle to avoid being identified, told Motherboard. “For biology, we learn about biotech and write five good and bad things about biotech. I send the AI ​​a prompt, ‘What are five good and bad things about biotech?’ And that creates an answer that gets me an A.”

Without AI, Innovate_Ry says homework that they consider “busy work” would take two hours. Now homework assignments like this take 20 minutes.

“I like to learn a lot [and] “Sometimes the schoolwork I’ve done before makes me procrastinate and not turn in an assignment,” explains innovate_rye. “Being able to do it faster and more efficiently is like a skill for me.”

innovate_rye is not alone. Ever since OpenAI unveiled the latest application programming interface (API) for its widely used language model, GPT-3, more and more students have started providing written prompts. OpenAI’s Playground and similar programs that use deep learning to generate text. The results continue the initial prompt in a natural-sounding way and Often indistinguishable from human-written text.

When AeUsako_ was a high school senior last spring, they used OpenAI to create an entire essay about contemporary world affairs. They told Motherboard that even though they didn’t ace the assignment—they lost points for failing to cite outside sources—they learned that plagiarism-checking algorithms don’t flag AI-generated text.

“Because I used Open AI, I didn’t have the constant worry of having to focus my time on writing,” AeUsako_, who also asked to use her online pseudonym, told Motherboard.

George Veletsianos, Canada Research Chair in Innovative Learning & Technology and associate professor at Royal Roads University, says text generated by systems like the OpenAI API are technically the original outputs of a blackbox algorithm.

“[The text] It’s not copied anywhere else, it’s produced by a machine, so plagiarism checking software can’t detect it and pick it up because the text isn’t copied anywhere else,” Veletsianos told Motherboard. “Without knowing how these other plagiarism checking tools work and how they will be developed in the future, I don’t think AI text will be recognized that way.”

It is unclear whether the companies behind the AI ​​tools have the ability to detect or prevent students from using them to do their homework. OpenAI did not comment at the time of publication.

Peter Laffin is a writing instructor and founder of the private tutoring program Crush the College Essay. He says tools like OpenAI are emblematic of other compensation methods that technology has produced over the past decade, such as cloud-based typing assistants.

“In literacy education, especially for emerging writers, instructors are looking for a desirable level of difficulty, or a point where you’re working hard enough to improve without breaking down,” Laffin told Motherboard. “Finding the right, appropriate level of difficulty will increase their ability to write. So if you’re finding that level of difficulty and doing compensatory techniques that go beyond providing instruction in that space, you’re not helping them grow as a writer.

Veletsianos notes that with AI-generated text we are likely to have passed the point of no return and students are not the only ones being treated.

“We can also start to see where this technology generates all kinds of questions around discourse and discourse on the fly,” he said. “I’m not saying the system we have is the best system, but I’m saying these are conversations we need to have and look at how we can use these tools to improve not only the efficiency of instruction, but its effectiveness and engagement. Well.”

Although Laffin acknowledges that effective education needs a reevaluation, he says it can happen when educators look at the types of prompts they assign students, noting the difference between fact retrieval and information discovery. However, he worries that products like OpenAI’s text generator will make essay writing a more important topic.

“We miss the learning journey,” Laffin said. “We may know more but we’ve never learned how we got there. We’ve always said the process is the best part and we know it. The satisfaction is the best part. That might be the thing that separates it all. And I don’t know anyone who creates more than anything else. Beyond academics , I don’t know what a person is like if they never struggle through learning. I don’t know the behavioral implications.”

Meanwhile, innovate_rye is eagerly awaiting GPT-4 It is estimated that over 100 trillion machine learning parameters can be trained and can go beyond just text outputs.. In other words, they don’t plan to stop using AI to write articles anytime soon.

“I still do my homework on things I need to pass, I use AI to handle things I don’t want to do or that don’t make sense,” added innovate_rye. “If AI can do my homework now, what will the future look like? These questions excite me. ”



Source link