Against Tasks and Hallucinations: Returning to Thought in the Age of Machine Learning
Saved in:
| Title: | Against Tasks and Hallucinations: Returning to Thought in the Age of Machine Learning |
|---|---|
| Language: | English |
| Authors: | Lily Lucas Hodges |
| Source: | History Teacher. 2025 58(4):409-430. |
| Availability: | Society for History Education. California State University, Long Beach, 1250 Bellflower Boulevard, Long Beach, CA 90840-1601. Tel: 562-985-2573; Fax: 562-985-5431; Web site: http://www.societyforhistoryeducation.org/ |
| Peer Reviewed: | Y |
| Page Count: | 22 |
| Publication Date: | 2025 |
| Document Type: | Journal Articles Reports - Evaluative |
| Education Level: | Higher Education Postsecondary Education |
| Descriptors: | Artificial Intelligence, Essays, Computer Uses in Education, History Instruction, College Students, Writing (Composition), Process Education |
| Geographic Terms: | California |
| ISSN: | 0018-2745 1945-2292 |
| Abstract: | A standard learning outcome for students in humanities courses, from middle school into higher education, is to craft clear, specific arguments based on evidence. A common assignment to develop and assess this learning outcome is the take-home essay. Generative AI models, such as ChatGPT, Copilot, and Grammarly, are quickly able to generate such essays, from the brainstorming process to the final draft. One way teachers assess if students used AI to write their essay is by running student papers through Turnitin, itself an example of AI. Since AI is not a word-for-word form of plagiarism, the best Turnitin can do is provide a measure of the probability of AI use. However, students have self-reported high usage of AI in surveys. Unsurprisingly, teacher concerns over student use of AI have increased. These concerns seem well-founded, considering a 2023 study revealed that a majority of students indicated they would continue to use AI even if instructors prohibited it. In this article, the author reviews student discussions of AI and the take-home essay in a lower-division history course called "LGBTQ America" at Chapman University in California. After the discussions, students then investigated AI-generated essays through a "deconstructed essay" and reflected on the process. An analysis of their class discussion and submitted work offers insight into the limitations and possibilities of engaging with AI, highlighting both a general student desire to continue writing take-home essays and a unique set of teachable moments offered by engagement with generative AI. |
| Abstractor: | ERIC |
| Entry Date: | 2025 |
| Access URL: | https://www.societyforhistoryeducation.org/A25Preview.html |
| Accession Number: | EJ1482241 |
| Database: | ERIC |
| Abstract: | A standard learning outcome for students in humanities courses, from middle school into higher education, is to craft clear, specific arguments based on evidence. A common assignment to develop and assess this learning outcome is the take-home essay. Generative AI models, such as ChatGPT, Copilot, and Grammarly, are quickly able to generate such essays, from the brainstorming process to the final draft. One way teachers assess if students used AI to write their essay is by running student papers through Turnitin, itself an example of AI. Since AI is not a word-for-word form of plagiarism, the best Turnitin can do is provide a measure of the probability of AI use. However, students have self-reported high usage of AI in surveys. Unsurprisingly, teacher concerns over student use of AI have increased. These concerns seem well-founded, considering a 2023 study revealed that a majority of students indicated they would continue to use AI even if instructors prohibited it. In this article, the author reviews student discussions of AI and the take-home essay in a lower-division history course called "LGBTQ America" at Chapman University in California. After the discussions, students then investigated AI-generated essays through a "deconstructed essay" and reflected on the process. An analysis of their class discussion and submitted work offers insight into the limitations and possibilities of engaging with AI, highlighting both a general student desire to continue writing take-home essays and a unique set of teachable moments offered by engagement with generative AI. |
|---|---|
| ISSN: | 0018-2745 1945-2292 |
