The End of the Final Answer
For generations, the final submission was king. A polished essay, a correct calculation, a completed project—these were the artifacts educators used to measure understanding. But the arrival of generative AI has changed the game entirely. Students now
have tools that can produce well-written essays, solve complex equations, and even write computer code in seconds. This has exposed a fundamental weakness in our traditional assessment models: they were designed for a world where producing the final answer required doing the work. Now, a flawless submission provides little guarantee that the student has genuinely engaged with the material. The challenge for educators is no longer just about preventing cheating, but about questioning whether the old ways of measuring knowledge are still fit for purpose.
Why Banning AI Isn't the Answer
The initial reaction from many institutions was to try and police AI use, leading to an unwinnable surveillance arms race with detection software. However, these tools are often inaccurate, create an adversarial relationship between teachers and students, and fundamentally miss the point. Banning these tools is futile when they are readily available on every student's phone and will be integral to many future workplaces. The more effective and forward-thinking approach is not to fight the technology, but to change the pedagogy. Educators are realising that this moment offers a powerful opportunity to reform assessment practices that were, in many cases, already poor proxies for deep learning. The goal is to prepare graduates who can work alongside AI, not compete with it.
Making the Process the Product
So, what does 'evidence of thinking' look like in practice? It means shifting the focus from the what to the how. Instead of only grading a final paper, educators can implement process-oriented assessments. This can include incremental assignments where students submit drafts, outlines, or annotated bibliographies, receiving feedback along the way. Other effective methods include in-class discussions, oral defenses where students must explain their reasoning and respond to questions, or assignments that require students to critique an AI-generated response, identifying its biases and inaccuracies. Some educators are using 'role-play assessments' that place students in realistic scenarios where they must make decisions and justify them, a task that can't be outsourced to a generic AI prompt. The common thread is making the student's learning journey visible.
Building Skills for a New Era
This shift toward process-based assessment does more than just make learning AI-resistant; it builds more durable and relevant skills. By focusing on the journey, students develop critical thinking, problem-solving, and resilience. They learn how to formulate good questions, evaluate information, and articulate their reasoning—abilities that are becoming more valuable than rote memorization. This approach also promotes 'AI literacy', teaching students how to use these powerful tools responsibly and ethically. It encourages them to see AI as a collaborator for brainstorming or research, not a replacement for their own intellect. Moreover, breaking down large assignments into smaller, manageable steps can reduce student stress and spread out the workload for both students and staff.
