AI bot ChatGPT gives unscrupulous students a new way to cheat on essays

Artificial intelligence is now generating its own content for the first time on a wide scale. ChatGPT is also generating some controversy, especially on college campuses.

The AI application is only 49 days old, but it has a massive vocabulary and knows most history before 2022.

It’s mainly used for writing just about anything.

ChatGPT can tell you the top 5 pho restaurants, then write computer code and an essay about Alan Turing — all in just a few seconds.

"It is a conversational model because it is conversing with you," said University of St. Thomas software and data science professor Dr. Manjeet Rege.

He says ChatGPT is a chatbot, like you might’ve interacted with for customer service online.

But it’s very advanced and barely leaves a mark revealing its work was performed by artificial intelligence.

Some University of Minnesota students told us they know people who’ve used it to write papers or help with homework.

"It's just interesting in this day and age to see how far technology has progressed and the lengths people will go to just not write the essay," said UM student Celine Sackih.

"I definitely think professors will get smarter about it and maybe use harder questions, but I don't know," said a UM student named Hannah. "In the end, the professors know there's a lot of resources to commit plagiarism, I guess."

Dr. Rege says researchers are already at work to detect cheating with ChatGPT.

Its makers are working on techniques to watermark what it writes.

And while its writing may not always be recognizable to humans, other artificial intelligence may be able to notice.

"When a machine or an AI model generates content, there is a particular pattern and that pattern occurs at particular intervals, so there's a particular pattern that you can detect that this has not been generated by a human," he said.

Critics worry ChatGPT could make a lot of jobs obsolete, especially writers like news reporters, but Dr. Rege says it’ll always need human handlers to create a perfect product.

ChatGPT writes a news article

To wrap up this story, we asked it to describe the future of ChatGPT.

"The future of ChatGPT will likely involve continued improvements in its ability to understand and respond to natural language input," it wrote.

We also asked it to write an entire news story about educators trying to prevent cheating with ChatGPT and this is what it generated:

Educators across the country are taking steps to prevent students from using advanced language models like ChatGPT to cheat on exams and assignments. In recent years, the rise of these powerful AI-based tools has made it easier for students to quickly generate high-quality written work, raising concerns that some may be using them to plagiarize or pass off work as their own.

To combat this problem, educators are implementing new measures to detect and prevent cheating, such as using plagiarism-detection software and requiring students to submit work in a format that makes it more difficult to copy and paste text. Some are also incorporating more open-ended, creative assignments that are less susceptible to cheating.

In addition to these technical solutions, educators are also working to educate students about the importance of academic integrity and the consequences of cheating. Many schools are incorporating lessons on digital literacy and responsible use of technology into their curriculum, to help students understand the potential risks and ethical implications of using advanced language models like ChatGPT.

Despite these efforts, some experts warn that it may be difficult to completely eliminate the use of ChatGPT for cheating, as it can be hard to detect and difficult to block without severely restricting students' access to educational technology. However, educators and educational institutions are making efforts to strike a balance between providing students with access to these powerful tools and ensuring that they are used ethically and responsibly.