University professors across the United States are confronting a new wave of academic dishonesty fueled by large‑language‑model (LLM) systems such as ChatGPT. Small, formative quizzes that once served as low‑stakes study aids now risk being completed in seconds by an "agentic" AI browser, prompting a debate over whether to preserve these assignments for honest learners or discard them to curb cheating.

Many instructors have reverted to traditional evaluation methods that resist AI manipulation. Oral examinations, supervised handwritten essays, and in‑person testing remain the most reliable safeguards. However, those solutions are largely unavailable to teachers of asynchronous online courses, a format that has become essential for students with physical disabilities, learners in remote regions, and adults balancing work or caregiving responsibilities.

"If we have to simply give up on the idea of online classes, those are the casualties," one educator warned, underscoring the stakes for a population that depends on digital learning. The challenge is not merely technical; it strikes at the heart of pedagogical quality. Oral exams, while effective, are labor‑intensive and have never been threatened by rising student‑to‑instructor ratios. In contrast, pen‑and‑paper or keyboard‑based tests, once praised for consistency and reduced scoring bias, now sit vulnerable to mass automation.

Writing assignments, once prized for fostering critical thinking, are among the first to be cut. A natural‑disasters professor recalled assigning students to script a Hollywood‑style disaster film, a task that blended accurate physics with creative storytelling. The exercise sharpened writing skills and encouraged deep application of course concepts. Today, such assignments risk becoming obsolete if AI can produce comparable drafts with minimal effort.

Educators are exploring hybrid approaches. Some propose redesigning quizzes to require real‑time data entry, personalized reflection, or multi‑modal responses that AI cannot easily replicate. Others suggest integrating AI as a learning tool rather than a threat, allowing students to interact with language models under supervised conditions to develop prompt‑engineering skills.

Nevertheless, the core tension remains: safeguarding academic integrity without sacrificing the flexibility that makes online education viable. As institutions weigh options, the fate of formative assessments hangs in the balance, and the broader conversation about AI's role in higher education continues to evolve.

Used: News Factory APP - news discovery and automation - ChatGPT for Business