OpenAI, an artificial intelligence company, has created a free online tool known as ChatGPT, a language processing program that uses AI to generate responses to designated prompts.
College students have been using ChatGPT as a tool for their assignments to explore different ways of responding to prompts for written assignments.
QuickBites, an initiative sponsored by San Jose State’s Humanities and Arts in Action program, held “ChatGPT: Help or Hype?” a virtual event where professionals from SJSU discussed this topic.
The event was hosted by Sara West, assistant professor of professional and technical writing, and education professor Roxana Marachi.
“If students are going to be using it, they are cheating themselves, they’re not getting the education they paid for,” Marachi said.
West and Marachi spoke about AI and what the existence of ChatGPT means for the future of students' learning.
ChatGPT accesses information from the internet to retrieve and apply knowledge to the written response it generates.
Marachi said not all the information from the AI is accurate and that the AI is a “con artist.”
“ChatGPT is really good at telling us what we want to hear,” Marachi said. “Someone who is struggling may ask ChatGPT what to do, and that is a problem.”
West and Marachi tested the accuracy of the program by putting in prompts relating to their research.
They found that ChatGPT would have false information, citing things that were incorrect.
“It's designed to almost produce that misinformation, because it's not designed to go look up the person and then give you the information,” West said. “It’s designed to guess what word is next.”
During the event, the speakers discussed how ChatGPT has the ability to generate written responses to prompts based on data patterns.
West and Marachi said there is no guarantee the information will be completely accurate and there have been instances when ChatGPT has failed in generating factual information.
There was also a discussion on the AI potentially inputting harmful, discriminatory information that could create a bad reputation for the student if they're not careful.
“Then you’re going to be responsible for that text, not the AI,” West said.
The Peabody EDI Office of Equity, Diversity and Inclusion from Vanderbilt University sent out an email to the school population regarding the recent Michigan State University mass shooting that occurred on Feb. 13.
The office used ChatGPT to generate the email and students were outraged by the misinformation in the email and the lack of human thoughts that went into the public statement, according to Vanderbilt’s school newspaper, the Vanderbilt Hustler.
According to West and Marachi, students must be wary when using this program for writing their papers. It can take away from the learning process and it may not answer the prompt as needed.
Professors are aware of ChatGPT, and they are cracking down on assigning students prompts that will prove if a student is using the program or not.
English senior Fatima Bello said she believes that ChatGPT would hinder students and their writing.
“I feel like it would not help students, many students are already struggling with writing as it is,” Bello said. “It can definitely help guide them, but many students would not use it that way. They would use it to write their research papers and they wouldn’t learn how to write.”
West and Marachi talked about the dangers and benefits of ChatGPT.
They discussed how it can be a method to help point students in the right direction of what to write, leading them to sources they can explore in the writing process, but they also mentioned why ChatGPT has the potential to be harmful.
