Controversy over using AI to score

Controversy over using AI to score Picture 1

When receiving essays from students, strategic communications professor Diane Gayeski (Ithaca College) will enter a portion of content into ChatGPT and then ask the system to criticize and suggest ways to improve.

'The best way to use AI for grading is to let them act as teaching assistants or research assistants in charge of looking over the first essay. AI does the job quite well,' according to Professor Gayeski.

She showed students the feedback from ChatGPT along with the corrections written by the system, and also shared her thoughts. Then they discussed together.

Professor Gayeski asked students to do something like importing their thesis draft into ChatGPT to see which content could be written better.

AI is making education change drastically. New technology brings practical benefits such as automating some tasks to help focus time on teaching, but at the same time there are risks regarding the accuracy of content provided or plagiarism.

Both teachers and learners are using AI. Teachers use a variety of tools such as ChatGPT, Writable, Grammarly, EssayGrader for grading, lesson planning, writing exercises, polls, creating videos or interactive activities. Students rely on ChatGPT or Microsoft CoPilot (integrated in the Microsoft Office software suite) to search for documents, summarize the content of documents to read or support essay correction.

Some schools are starting to develop policies on how students use AI, but there are no guidelines for teachers. The use of AI to grade or critique assignments also raises controversy about academic ethics. Both parents and students may question whether AI scores and comments are worth the money and time they spend.

Professor Gayeski recommends that teachers rely on AI to consider certain factors such as layout, grammar, and word usage, then grade them themselves. They should also consider the depth of knowledge and creativity in students' essays.

Dr. Leslie Layne (University of Lynchburg) admits that AI has both benefits and harms. She commented that importing part of a student's essay content into ChatGPT risks violating intellectual property rights. This is the data source to train a series of generative AI systems.

According to her, AI should especially be avoided with master's theses and doctoral theses because students may hope to publish their articles. The female doctor emphasized: 'It would be wrong to enter the essay into AI without letting the student know first. Consent from the student is required'.

Some teachers use Writable software (based on ChatGPT) for grading. The software has encryption feature so the essay does not contain personal information and the information is not shared directly with the system. Many other platforms like Turnitin support determining whether an essay is written by a human or an AI, but such tools are not necessarily accurate.

Several schools are actively building policies on using AI for both teachers and learners, one of which is Johns Hopkins University. Dr. Alan Reid - a member of the AI ​​policy drafting committee of Johns Hopkins University - said that AI systems are currently used by many schools to prepare promotion or appointment documents, evaluate work performance and Published Recruitment.

Associate professor of philosophy Nicolas Frank (Lynchburg University) is concerned that it is still too early to know how widely AI will be applied, plus some educational managers who do not participate in teaching will build it. The policy is too simple. He recommends first identifying a range of AI abuses and then formulating policy based on this.

Ms. Leidner believes that universities can provide general guidance, such as prioritizing transparency (telling students in which cases AI is used to grade) or determining what types of data are not allowed to be provided. granted to AI.

Samuel Daniel

Update 09 April 2024