The AI Cheating Arms Race: Why Watermarks Won’t Save Education
The education world is in a panic. AI tools like ChatGPT and Claude have become the ultimate study buddies—or, depending on your perspective, the ultimate cheating enablers. Now, companies like Anthropic are rolling out AI-watermarking systems, embedding invisible markers in AI-generated text. On the surface, it sounds like a silver bullet: teachers can finally catch those sneaky students red-handed. But personally, I think this is a classic case of treating the symptom, not the disease.
The Illusion of Control
Let’s start with the watermark itself. Anthropic’s Claude now includes a hidden label in its outputs, and OpenAI has a similar tool in the works. What many people don’t realize is that these watermarks are far from foolproof. They can be stripped away by other AI tools, and they don’t account for AI-assisted editing or proofreading. In my opinion, this is like putting a “wet paint” sign on a fence—it might deter some, but it won’t stop the determined.
What this really suggests is that we’re in an arms race. Students will find ways to bypass watermarks, and developers will create new detection methods. It’s a never-ending game of cat and mouse. If you take a step back and think about it, this cycle is exhausting and ultimately unproductive.
The Bigger Problem: AI Isn’t Going Away
Here’s the thing: AI is here to stay. A 2026 UK survey found that 95% of students use AI tools, with 94% incorporating them into their assessments. Banning or detecting AI use feels like trying to stop the tide with a spoon. From my perspective, the real issue isn’t whether students are using AI—it’s how we’re teaching and assessing in an AI-driven world.
One thing that immediately stands out is the disconnect between educators and reality. Many institutions are still clinging to outdated assessment methods, like take-home essays, which are ripe for AI abuse. New South Wales’ decision to ban take-home assignments for Year 12 students is a step, but it’s reactive, not proactive. What we need is a fundamental rethink of how we evaluate learning.
The Shift We Need
Educational psychologist Jason Lodge argues that we should stop focusing on whether students are cheating and start focusing on what they’ve learned. This resonates deeply with me. In an era where AI can write essays, solve equations, and even debug code, traditional assessments are becoming obsolete.
A detail that I find especially interesting is the push for “secure” assessments—oral presentations, in-class tasks, and supervised demonstrations. These methods not only verify the student’s identity but also test their ability to apply knowledge in real-time. However, they’re not a one-size-fits-all solution. For online learners, in-person assessments are impractical. And let’s be honest, oral exams come with their own biases and challenges.
The Ethical Quandary
Another layer to this debate is the ethical use of AI. OpenAI reportedly withheld its watermarking tool due to concerns about its impact on non-native English speakers. This raises a deeper question: Are we creating systems that unfairly penalize certain groups? AI detection tools have already faced legal challenges over accuracy and bias. If we’re not careful, we risk exacerbating inequalities in education.
The Future of Learning
So, where do we go from here? Personally, I think the answer lies in embracing AI as a tool for learning, not just a threat to it. Assessments should be designed to encourage critical thinking, creativity, and ethical AI use. For example, why not teach students how to evaluate AI-generated content or collaborate with AI to solve complex problems?
What makes this particularly fascinating is the potential for AI to transform education entirely. Imagine a world where AI tutors personalize learning, and assessments are adaptive, reflecting a student’s growth over time. This isn’t science fiction—it’s already happening in pockets around the globe.
Final Thoughts
AI watermarking might seem like a solution, but it’s a band-aid on a bullet wound. The real challenge is reimagining education for an AI-driven future. In my opinion, the focus should be on fostering skills that AI can’t replicate—creativity, empathy, and ethical judgment.
If you take a step back and think about it, the rise of AI isn’t a crisis—it’s an opportunity. The question is, will we seize it, or will we keep chasing our tails in the detection game?
