Eight Filipino instructors and administrators reflected on their own AI use: shared gains, shared worries about cheating, and no institution with a full policy

What benefits, risks and limits do instructors and administrators in the Philippines see in generative AI (ChatGPT, Bing, Quillbot, Grammarly) for teaching, administration and research, and what have their institutions done, in the first months after ChatGPT's launch?

Collective autoethnography between March and May 2023: each author wrote a reflection on their own experience and the group compared them. Participants had to hold a master's or doctoral degree, be affiliated with a Philippine higher-education institution, and have used AI tools in research for more than eight months. Eight reflections from four institutions; one of the authors is a researcher at the centre. The journal published it as a column.

• Uses reported across the essays: writing quizzes and converting question types; generating ideas and simplifying difficult concepts; designing activities and rubrics; outlining; correcting grammar and tone, which matters most for writers whose first language is not English; synthesising literature; finding patterns in large bodies of text; paraphrasing.
• Limits met first-hand: factual errors, fabricated references, a knowledge cutoff, missing context and personal insight; students submitting AI essays as their own or copying answers without understanding and then failing oral checks.
• Shared worries: cheating, fabricated data, a decline in creativity and critical thinking, skill atrophy like the calculator user who can no longer do arithmetic, and questions of authorship and intellectual property. One author describes refusing a colleague's AI-written section of a joint manuscript.
• Attitudes ranged from “optimistic but not fanatic” to hesitant. The centre's researcher told students plainly that he could not detect AI use, so he required them to edit, personalise and verify anything AI produced, and worried that handing all the work to tools would limit students' critical thinking.
• None of the four institutions had a full generative-AI policy at the time. The measures in place were existing student codes of conduct, AI detection switched on in the plagiarism checker, orientation sessions and webinars, and policy drafting under way. The authors agree that policies and ethical guidelines are needed for students, instructors and administrators alike, and that prohibition would be futile.

A snapshot of how faculty actually used the tools in the first months after ChatGPT, useful for Thai faculties drafting AI policy. The practical line the authors propose is between “assistive” use (grammar, reference management, paraphrasing) and “generative” use (producing the answer), combined with transparency about what was used.