Teachers Fear ChatGPT Will Make Cheating Easier Than Ever
“There’s a lot of cheap knowledge out there. I think this could be a danger in education, and it’s not good for kids,” said one educator of OpenAI’s viral chatbot.
penAI’s latest chatbot, ChatGPT, can write elaborate essays and movie scripts, debug code and solve complex math problems. Its ability to generate legible answers to any question you can imagine could be a promising supplemental resource in classrooms, especially given the national shortage of teachers. But teachers are concerned about students using the free and accessible tool as a Wikipedia replacement to complete homework and to write assignments for them, endangering students’ willingness to develop skills like writing and researching.
“Students are going to think and use this chatbot as if it is a know-all,” says Austin Ambrose, a middle school teacher in Idaho. “That’s because it’s a technology that is creating these things that sound really legitimate, they are going to assume that it is and take it at face value.”
Over the past few weeks, ChatGPT has exploded in usage, with more than a million users signing up to use it within a week after it was launched. The algorithm is a language model trained through human feedback and a vast amount of public data from a variety of sources like books and articles from the internet. But just because it appears to know what it is talking about doesn’t mean that the information it provides is fully accurate. For one, because it was trained on data available until 2021, ChatGPT isn’t able to provide factual up-to-date answers. It sometimes presents minor inaccuracies that it should know, given its training data — for instance, it said the the color of Royal Marines’ uniform during the Napoleon war was blue when it was actually red. In addition, ChatGPT struggles with confusingly worded questions, which could also lead to incorrect answers.
The algorithm also has bias problems, given that it was trained on vast amounts of data pulled from the internet. It can render racially biased content: When asked for a way to assess the security risk of travelers, it proposed some code that calculated risk scores, which spit out higher scores for Syrians, Iraqis, and Afghans than other airline travelers.
Open AI’s CEO Sam Altman himself acknowledged these pitfalls in a tweet, saying, “ChatGPT is incredibly limited but good enough at some things to create a misleading impression of greatness. It’s a mistake to be relying on it for anything important right now.”
With these concerns in mind, teachers say that it is even more important to teach digital literacy early on and emphasize the importance of critically assessing where information is coming from. Teachers say the tool could also emphasize and reinforce the use of citations in academic papers.
