Go Old School: Deterring AI Cheating on Exams
How restoring handwritten exams can stem AI cheating and protect integrity without sacrificing efficiency.
Over the past few weeks, I’ve come across a few articles discussing the prevalence of cheating on exams using conversational AI, like ChatGPT by students across the spectrum.
Some educators have all but surrendered to the technology all together. Other institutions have already reinstituted paper exams or are in the midst of reforming exam policies.
While these chatbots can be a great study aid (which I advocate), some students may be tempted to cheat and rely too heavily on them during exams. However, the answer is not to completely forbid the use of technology in classrooms.
Rather, schools can deter cheating by reverting to some of the traditional analog testing methods that served effectively for generations prior to the digital age.
By administering handwritten, proctored exams without devices, while still utilizing automated grading through optical character recognition, institutions can achieve a balanced approach. Testing conditions that disconnect students from the internet and digital assistants will limit the opportunity for high-tech cheating.
Yet analog exams graded efficiently via AI can allow teachers to maintain academic standards. With some reasonable constraints, schools can still promote educational technology while upholding academic honesty.
Temporary Restrictions
The simplest way to prevent students from consulting AI assistants during exams is to prohibit all digital devices from the testing environment. Schools should require students to completely power down and surrender any smartphones, tablets, smartwatches, laptops or other electronics before entering the exam room.
Possession of such gadgets should be grounds for disciplinary action. This reduces temptation and blocks easy avenues for students to communicate with chatbots. Proctors can help enforce the no-device policy by actively monitoring for contraband. While a device ban represents a big change in today’s tech-driven classrooms, it’s a necessary step to close the door on high-tech cheating.
Bringing Back Handwritten Exams
Along with taking away devices, tests should be written by hand with old-fashioned pens and paper.
When you type answers to essays on a keyboard, it is easy to cheat with AI features like autocomplete and text generation. But if students have to write their answers by hand, it is much harder for them to use bots.
Handwriting makes it harder to cheat because it requires fine motor skills. And unlike typed fonts, printed handwriting has individual marks that are hard to copy digitally. Teachers can give students the exact exam papers instead of letting them bring their own, which is another way to discourage bad behavior. The tactile, hand-crafted nature of analog tests makes them far more secure than digital exams.
Here is a draft of the fourth and fifth sections for the blog post outline on preventing AI cheating through reverting to analog exam formats:
Vigilant Proctoring
Even if an exam is written by hand and has no devices, schools still need proctors to keep cheating from happening. Teachers or designated staff should actively monitor the testing environment to ensure policy compliance.
Proctors should keep an eye out for students trying to sneak in digital devices or materials that are not allowed. Active patrols around the room can also stop students from working together or sharing answers. Sharp proctoring helps catch real-time cheating efforts. Even if exam rules are written down, they still need to be enforced by a person.
Automated Grading of Analog Tests
The manual grading of submitted analog exams can seem like a daunting task. But luckily, modern technology is here to help.
Multiple choice sections can be graded automatically by feeding handwritten exam sheets through optical character recognition software. For essay and long-form questions, teachers can still provide individualized feedback and scoring.
This balanced approach applies automation to simplify routine grading, while preserving human evaluation of complex responses. Students get the integrity boost of analog tests, without creating major bottlenecks in the grading process.
Everything here is completely scalable.
Benefits
Beyond just deterring cheating, administering old-fashioned handwritten exams has pedagogical advantages. The process of handwriting engages different cognitive skills and motor pathways than typing. Students must break arguments and concepts down internally before reformulating them clearly in their own words on paper.
It forces clearer critical thinking to arrange thoughts without constant digital revision. Sitting exams by hand also provides practice in retaining key information without reliance on devices, improving knowledge internalization.
Rather than being a relic of the past, handwritten essays and test questions stimulate learning processes that remain vital to intellectual development. That is why schools should embrace handwriting exams not just for security but for improving true academic skills.
The Takeaway
It will take creativity to deal with new ways that AI could be used to cheat while still promoting the benefits of technology in education. Instead of a complete ban on tech in the classroom, a more balanced solution would be to use tightly supervised analog exams without devices and AI grading systems.
With this well-thought-out compromise, we can discourage bad behavior and protect academic integrity without sacrificing efficiency. By using the best parts of both analog and digital methods, institutions can make it impossible for even the most clever student to cheat. Some people might think that tests that have to be written by hand and can not be taken on a device are a step backwards in terms of technology. But keeping values like honesty and hard work is also part of progress.
With care, schools can use the best of both old and new ways of teaching to keep education moral and effective in the age of AI.
What do you think?
I’m interested in topics and trends that intersect cognitive psychology, meta-learning, behavioral economics, and technology. I also run a venture focused on innovation and commercialization of services that lifts humankind and heals our planet. If you have a big idea, visit me at signetscience.com .
