Shin Ji Won reports
Editor: Editorial Office
ChatGPT has begun to rage in universities around the world. Recently, HKU fired the first shot of the domestic AI counterattack war.
As AI-generated text becomes more and more similar to human writing, high school and college teachers are drowning in papers suspected of AI cheating.
The most frightening thing is that it is difficult to distinguish between real and fake papers, which greatly increases the workload of teachers.
Recently, the University of Hong Kong issued a notice to all teachers and students that the use of ChatGPT or other AI tools in all classes, assignments and assessments is prohibited, otherwise it will be regarded as "plagiarism".
HKU also became the first university in Hong Kong to explicitly ban the use of ChatGPT.
In this regard, Hong Kong Chinese University said that it will meet next week to review the situation and formulate applicable policies; While Hong Kong Baptist University does not explicitly prohibit the use of ChatGPT, it emphasizes the utmost importance to the academic integrity of students and teaching staff.
HKU prohibits the use of ChatGPT
According to a number of Hong Kong media reports, Ian Holliday, vice president of the University of Hong Kong, sent a letter to teachers and students on February 17, asking that the use of ChatGPT or other AI tools to take classes, do homework or exams be prohibited on campus.
If necessary, prior written permission of the relevant course instructor is required, and violations of the above temporary measures will be considered "potential plagiarism"; If teachers suspect that students are using ChatGPT, they can ask students to discuss relevant papers or works, set up additional supplementary oral examinations, add classroom exams, etc.
Ho explained that as ChatGPT takes the internet by storm and shakes up global education, HKU plans to launch a broad campus debate on the impact of AI on school teaching.
"Because of the impact, it will take us a while to define a long-term policy," he said.
"Therefore, we need to adopt a short-term policy: as a temporary measure, we prohibit the use of ChatGPT or any other AI-based tool in all classrooms, courses and assessment tasks at HKU."
Ho warned that violations of the temporary policy would be considered plagiarism, and said teachers could use a variety of methods to verify that students were aided by artificial intelligence technology.
"Teachers who suspect the use of ChatGPT or other AI-based tools can ask students to discuss their results, supplementing oral exams, offline exams or other actions if necessary," he said.
Casey Chik Yau-hong, an undergraduate representative at the University of Hong Kong, said the ChatGPT ban was acceptable as a stopgap measure. He added that the technology is unfair to many students.
In his opinion, how to detect the content of artificial intelligence is more worth thinking. Chik said he was more focused on the promised outcome of the consultations.
"In the longer term, I'm more focused on how universities will provide the space for students to make the most of these tools, including how to adapt to the technology," Chik said.
Hong Kong School Joins AI Strangling Brigade
Hong Kong Baptist University said in a statement that it expects students to make appropriate use of technology in their academic pursuits and creations, but any claim that a non-original work belongs to their own work will be disciplined for plagiarism.
However, Baptist University did not specify whether the use of AI-assisted tools would be considered plagiarism.
City University of Hong Kong said the university had not yet made a decision on the issue because the assessment would take time.
Meanwhile, HKUST said it has developed procedures to deal with plagiarism and will closely monitor the technology's impact on academic and research efforts before reviewing its policies.
According to SCMP, Francis Fong Po-kiu, honorary president of the Hong Kong Information Technology Association (HKITF), there are various tools to detect if students are using any AI-powered tools to write essays or do coursework.
Fong Po Kiao, Honorary President of the Hong Kong Information Technology Association
"If students rely heavily on AI-based tools to complete the curriculum, they will not be able to develop their logical reasoning, critical thinking and language skills. They will lose the independent process of learning and verification of knowledge," he said.
In Dutch universities, the use of text written by AI is classified as academic fraud, but many academic institutions say how to judge ChatGPT content is a difficult point for censorship.
Since the launch of ChatGPT, education experts have conducted research on its testing capabilities. They said they could get perfect marks for the course if they submitted ChatGPT-generated answers by undergraduates.
For programmers, ChatGPT is more like a god. After all, coding challenges can be solved even in seconds with the tool.
But ChatGPT's inability to answer users' questions by tapping into the vast amount of text on the internet and making probability-based guesses means it still can't solve queries based on real-time information and can provide reasonable but incorrect answers.
OpenAI says that ChatGPT is currently not used in areas where high accuracy is required, and is more suitable for tasks that involve calling facts, rather than tasks that require "logical reasoning."
Foreign students have set off a wave of ChatGPT cheating
before
The news that a student at a university in the United States used a paper written by ChatGPT to win the first place in the class made people exclaim whether the education system was really going to be subverted by AI
。
A Jan. 24 poll by Nature found that teachers are rethinking homework formats to address the surge in cheating papers.
These teachers didn't worry about it.
According to one survey, 89% of American college students are already using ChatGPT for homework
。
More recently, the ChatGPT cheating scandal shocked the school by erupting in an elite academic program at a Florida high school.
The teacher was distressed: Why did such a smart student choose to cheat?
Recently, the project's leaders discovered that some students submitted papers that were clearly problematic — completely different from the style of their previous papers.
The ChatGPT cheating scandal rocked Cape Coral High School in Florida
Several of the students have admitted to using ChatGPT themselves. But some students did not admit that the person in charge had planned to confront the suspected student.
She stressed that if the school later found out that these students had indeed misbehaved, the consequences they faced would be even more serious.
The person in charge said that students who cheated will be disqualified from graduation.
The scandal has shocked the whole school, because this elite program recruits outstanding students.
One teacher said, "It's heartbreaking that we have the smartest, hardest working and most competitive kids and that they choose to cheat. The only good thing is that only a few people are cheating with AI at the moment."
Teachers at Cape Coral High School are discussing in emails which students are suspected of cheating with AI
Now, teachers hope that the fear of cheating being discovered, and the possible harsh punishment, can serve as some deterrent to students.
However, the difficulty in detecting ChatGPT cheating is that it generates a different language each time, so traditional plagiarism detection procedures are ineffective.
To that end, school leaders are analyzing students' Chromebooks to find out which assignments are suspicious. In addition, we take measures such as regular interviews with students to check their knowledge level.
Teachers believe that this approach ensures that students' work "reflects what they understand."
In fact, in order to curb the constant fear of cheating,
OpenAI, maker of ChatGPT, has introduced tools to detect AI-generated work
, but the accuracy rate is not guaranteed.
Experience address: https://platform.openai.com/ai-text-classifier
Several major school districts, including New York City, Los Angeles, Seattle, and Virginia, have long disabled AI such as ChatGPT on school devices and networks.
The head of the Elite Project said, "We will not tolerate cheating. Any student who violates the Code of Conduct and Academic Integrity Policy will be disciplined."
"As part of their cybersecurity efforts, the school's information services team will continue to strengthen the security features of Chromebooks to discourage students from using AI to complete their assignments."
In AI ethics classes, use AI to cheat
Coincidentally, according to NBC Bay Area, in an "AI ethics" class, students turned in papers written in ChatGPT.
"It's ironic," said Professor Brian Green, head of the Ethics of Technology program at Santa Clara University. "This article is not entirely on topic, and its wording has a 'robot flavor'."
Therefore, Professor Green, who was crying and laughing, simply changed the format of the final assignment: it would no longer be in the form of a thesis, but the students themselves would give oral presentations.
Brian Green
According to ABC, several Bay Area universities have convened all professors to discuss how to deal with the impact of ChatGPT on teaching in schools. And these news are just the latest of many ChatGPT cheating cases.
Of course, the school's fear of ChatGPT has also caught the attention of maker OpenAI. CEO Sam Altman said in a recent interview that schools need to overcome it.
"Generative text is something we all need to get used to," Altman said. "We adapted to calculators and changed the tests we tested in math class. There is no doubt that ChatGPT is a more extreme version, but at the same time its benefits are more extreme."
It's true that ChatGPT is much more powerful than calculators and more useful for students who like to cheat.
At least one point, Altman is right — in a way, we are living at the beginning of a new era.
Many basic writing tasks can be completed by AI, and we only need to do very little human input, especially some composition topics taught in high school English classes.
Now, the pressure is on the teacher's side
But it has to be said that the use of ChatGPT in the classroom of AI ethics is indeed somewhat ironic.
In response, Professor Green said that our purpose was to train engineers, writers, businessmen, etc. for society, only to find that they only cheat in all courses.
And this involves some very basic questions about what the education system does, how it works, and what role it should play in society.
"If we can't assess whether these students are mastering what is being taught, we're in big trouble."
Unfortunately for teachers, ChatGPT and its successors will continue to grow.
If schools want to prepare students for life in the age of AI, they must be taught how to properly work with the latest chatbots.
Of course, this is not an easy task, after all, the teachers themselves have to figure out these AIs first.
Recall that when search engines first appeared, the question of "whether students using search engines to write homework is considered cheating" caused an uproar, and now, people have long been accustomed to "Baidu".
Perhaps, ChatGPT really just appeared too early.
Resources:
https://udn.com/news/story/7332/6980433
https://nypost.com/2023/02/16/chatgpt-cheating-scandal-erupts-at-florida-high-school/
https://gizmodo.com/ai-chatgpt-ethics-class-essay-cheating-bing-google-bard-1850129519