{"slug": "my-watch-was-a-calculator-education-survived", "title": "My Watch Was a Calculator. Education Survived", "summary": "Students will use AI for the same reason earlier generations embraced calculators: it makes work faster, but the scale of AI's capabilities poses new challenges for education, as shown by a Brown University exam where the class average was 96, leading to an in-person final that saw 27 students drop and the average fall to 48.6. Roberto Serrano, an economics professor at Brown, concluded widespread cheating after AI-generated answers were submitted, highlighting the need for education to distinguish between AI use that supports learning and AI use that substitutes for it.", "body_md": "TL;DR — Key Takeaways\n\n- Students will use AI for the same basic reason earlier generations embraced calculators, computers and search engines: It makes useful work faster and easier.\n- The Brown University exam case shows the danger of students submitting AI-generated answers without actually understanding the material.\n- AI use and AI cheating are not the same thing; the key distinction is whether AI\n**supports learning or substitutes for it**. - In-person tests, oral exams and classroom discussions can help verify independent knowledge, while other assignments can deliberately allow AI use.\n- Education needs to prepare students for a workplace where AI use will be expected, while ensuring they retain the subject knowledge needed to evaluate its output.\n- The goal should be\n**competent, transparent and accountable AI use**, not pretending students can be educated for a world without AI.\n\nI remember my first digital watch. You pressed a button and the time appeared in glowing red numbers. That alone seemed futuristic. Then came the [Casio calculator watch](https://www.casio.com/us/watches/50th/Heritage/1980s/), with its tiny LCD screen and buttons that were almost too small for a fingertip. I thought it was incredibly cool that I could do math on my wrist.\n\nI did not buy the watch to cheat, and it did not keep me from learning arithmetic. Something inside me simply said it would be silly not to use a tool that could help me calculate faster and more accurately. More than four decades later, students looking at artificial intelligence are reaching the same conclusion.\n\nThe difference, of course, is scale. My calculator watch could add, subtract, multiply and divide. Today’s AI systems can write essays, solve advanced mathematics, explain scientific concepts, generate software and complete assignments that once required hours of work. Comparing AI to a calculator can only take us so far, but the instinct to use the best available tool is the same. Education providers need to recognize that before they turn the presence of AI into an endless contest between increasingly capable students and increasingly elaborate methods of catching them.\n\nA recent [Washington Post article](https://www.washingtonpost.com/education/2026/07/15/even-elite-colleges-are-scrambling-root-out-ai-cheating/) provides a particularly dramatic example of the problem. Roberto Serrano, an economics professor at Brown University, became suspicious after seeing the results of a take-home midterm in his advanced mathematical economics course. The class average was 96. In previous years, it had ranged from the 60s to the 80s. Nearly half the students received a perfect score.\n\nOne problem produced an additional clue. When Serrano and his teaching assistants ran the exam through ChatGPT, it generated an odd, unnecessarily convoluted solution rather than a straightforward proof. Numerous students used a similar method.\n\nSerrano concluded that widespread cheating was the only plausible explanation. He converted the final into a three-hour, in-person examination. Twenty-seven students dropped the course, including 22 who had scored 100 on the midterm. The average on the final fell to 48.6. That is fairly persuasive evidence that something went badly wrong.\n\nStudents who violated explicit rules and submitted AI-generated answers as their own cheated. They represented those answers as evidence of knowledge they apparently did not possess. Nothing about integrating AI into education requires us to pretend otherwise.\n\nBut catching cheaters cannot be the entirety of an AI strategy for education. Take-home exam or not, students are going to use AI. They will use it to study, conduct research, write, code, solve problems and prepare presentations. When they graduate, their employers will expect them to use it. Schools can either teach students to use AI intelligently and responsibly or continue designing rules and assignments that reward them for hiding it.\n\n### What Are Tests Supposed to Measure?\n\nTests are important, but not because completing a test without assistance is inherently virtuous. Tests are supposed to measure what students have learned and how well they understand the curriculum.\n\nWe already recognize that access to information does not automatically invalidate an assessment. Open-book exams are not unusual. A well-designed open-book exam can be more demanding than a closed-book test because memorizing a definition is not enough. Students must understand the material, determine what information matters and apply it to a problem. The book contains information, but it does not demonstrate understanding on the student’s behalf.\n\nAI complicates that distinction because it can do more than retrieve information. It can synthesize material, construct arguments and generate answers. That makes it easier for a student to produce work without understanding it. It also makes AI a potentially powerful teaching tool.\n\nAn AI tutor can explain a concept one way and then try again when the explanation does not connect. It can generate practice problems, critique a student’s reasoning and provide feedback when a professor or teaching assistant is unavailable. A student who is reluctant to raise a hand in a lecture hall can ask AI the same question repeatedly without embarrassment. Someone struggling with a difficult passage can request a simpler explanation, an analogy or a practical example.\n\nIf that interaction helps the student develop a stronger grasp of the material, AI has not undermined education. It has supported it. The question should not simply be whether a student used AI. It should be whether the student learned.\n\n### Producing An Answer is Not the Same as Knowing\n\nThe Brown case demonstrates the difference between producing an answer and possessing knowledge. A student may be able to submit a perfect-looking solution generated by AI and remain unable to explain how it works. That answer is not evidence of mastery. It is evidence that the student has access to a tool. The challenge for educators is to design curricula and assessments that expose that difference.\n\nStudents should still learn foundational material. They should be able to explain their reasoning, defend their conclusions and apply concepts to unfamiliar problems. They should know enough to recognize when an AI system is hallucinating, relying on a faulty assumption or delivering a confident answer that happens to be wrong.\n\nSome of that may need to be tested in person. Oral examinations, classroom discussions and controlled assessments can help establish what a student understands independently. But education should not stop there. Other assignments should explicitly allow or even require AI use. In those cases, students could be evaluated on the quality of their questions, the judgment they apply to the output, the sources they use for verification and their ability to improve upon what the system produces.\n\nStudents might also be asked to defend an AI-assisted paper orally, identify weaknesses in an AI-generated argument or compare multiple AI answers and explain which is strongest. An engineering student could use AI to develop a solution and then be required to explain every design decision. A writing student could submit drafts showing how the work evolved, where AI contributed and why the student accepted or rejected its suggestions.\n\nThere will not be one assessment model for every subject or level of education. The objective is not to impose a fashionable layer of AI on every class. It is to determine what students need to know, what they need to be able to do and how educators can credibly establish that the learning occurred.\n\n### Stop Teaching Students to Hide the Tool\n\nCalculators produced their own educational panic. Teachers and parents worried that students would stop learning arithmetic because a machine could perform the calculations for them. Schools eventually found a workable balance. Students learned basic mathematics and then used calculators to tackle more advanced problems.\n\nThe internet produced another round of concern. Students suddenly had access to an enormous amount of information without visiting a library or opening an encyclopedia. Search engines and Wikipedia made finding answers easier, but they also made evaluating sources, distinguishing authority from popularity and synthesizing conflicting information more important. Education adapted because it had no realistic alternative. Nobody seriously believes that students should conduct research as though the internet does not exist. We teach them—or at least should teach them—how to use it critically.\n\nAI requires another adaptation. It will be more disruptive because the technology does not merely provide access to information. It participates in producing the work. That raises legitimate questions about authorship, attribution, accuracy and academic integrity. Those questions must be addressed.\n\nBut defining all AI use as cheating does not solve them. It teaches students to conceal the tool rather than master it. It encourages an arms race in which students find new ways to use AI invisibly while professors deploy unreliable detectors and more intrusive surveillance. Neither side gains much, and learning becomes collateral damage.\n\nEducation providers have a more demanding responsibility. They must make competent, transparent and accountable AI use part of the curriculum while preserving meaningful ways to test individual understanding.\n\nThat means distinguishing assistance from substitution. If AI helps a student understand a difficult concept, it is teaching. If it helps a student explore alternatives and examine their reasoning, it is teaching. If it produces an answer the student submits without understanding, it has become a substitute for learning. The technology may be the same. The educational outcome is not.\n\n### Education Has Adapted Before\n\nMy calculator watch did not destroy education. Neither did open-book exams, personal computers, the internet, search engines or Wikipedia. Each forced educators to reconsider what students needed to memorize, what they needed to understand and what they should be able to accomplish with the tools available to them.\n\nAI will force a larger reconsideration. It is capable enough to undermine assignments that once provided reasonable evidence of individual effort and understanding. The Brown exam makes that painfully clear.\n\nThe answer, however, cannot be to educate students for a world without AI. That world is already disappearing. Employers will not reward graduates for completing every task manually when better tools are available. They will reward people who can combine subject-matter knowledge with AI, exercise judgment over its output and take responsibility for the result.\n\nStudents will use AI because it would be silly not to. The job of education is to make sure they use it to learn. My watch was a calculator. Education survived. It will survive AI, too—but only if it is willing to adapt.", "url": "https://wpnews.pro/news/my-watch-was-a-calculator-education-survived", "canonical_source": "https://techstrong.ai/features/my-watch-was-a-calculator-education-survived/", "published_at": "2026-08-14 08:14:08+00:00", "updated_at": "2026-08-14 08:21:25.184495+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-policy", "ai-ethics"], "entities": ["Brown University", "Roberto Serrano", "Casio", "Washington Post"], "alternates": {"html": "https://wpnews.pro/news/my-watch-was-a-calculator-education-survived", "markdown": "https://wpnews.pro/news/my-watch-was-a-calculator-education-survived.md", "text": "https://wpnews.pro/news/my-watch-was-a-calculator-education-survived.txt", "jsonld": "https://wpnews.pro/news/my-watch-was-a-calculator-education-survived.jsonld"}}