최근 '아인슈타인(Einstein)'과 같은 에이전트형 AI가 등장하며, 온라인 강의 수강부터 퀴즈 풀이, 과제 제출까지 학생의 모든 학업 과정을 완벽하게 자동화할 수 있게 되었습니다. 이는 교육자들에게 거대한 혼란과 학업 부정행위에 대한 심각한 우려를 낳고 있습니다. 교육계는 AI가 교실을 완전히 장악하기 전에 이에 대한 현실적인 대응책과 제도적 논의를 서둘러 마련해야만 하는 중대한 기로에 섰습니다.
번역된 본문
윌리엄 류(William Liu)는 자신이 제때 고등학교를 졸업한 것에 감사하고 있습니다. 그는 최신 AI 도구가 그때 이미 있었다면, 숙제를 하는 데 그 도구들을 사용할 유혹을 느꼈을지도 모른다고 말했습니다. 현재 스탠퍼드 대학교 2학년인 류는 불과 2024년에 고등학교를 졸업했습니다. 그는 “두 살 어린 동생이 막 고등학교를 졸업했는데, 단 2년 차이일 뿐인데도 우리의 교육 경험은 완전히 달라졌다”고 말했습니다.
르가 졸업할 무렵에도 ChatGPT는 이미 교실에서 큰 혼란을 일으키고 있었습니다. 하지만 이제 학교 교육의 자동화는 더욱 심화되고 있습니다. 초기에 교사들이 학생들이 챗봇을 이용해 에세이를 쓰는 것을 걱정했다면, 이제는 '클로드 코드(Claude Code)'와 같은 새로운 에이전트 도구를 통해 학생들이 자신의 과제를 기계에 훨씬 더 많이 떠넘길 수 있게 되었습니다.
온라인 수학 퀴즈를 풀어야 하나요? 생물학 실험 보고서를 써야 하나요? 역사 수업용 파워포인트 프레젠테이션을 만들어야 하나요? AI는 이 모든 것과 그 이상을 해낼 수 있습니다. 한 고등학생은 최근 자신에게 AI가 대신 해주지 못할 과제를 생각해내는 것조차 힘들다고 말했습니다.
AI가 학교 과제를 얼마나 잘 해내는지 보여주는 척도로, '아인슈타인(Einstein)'이라는 새로운 봇을 생각해 보십시오. 몇 주 전, 이 도구는 엄청난 자랑과 함께 바이럴이 되었습니다. 이 봇을 광고하는 웹사이트는 “아인슈타인은 새로운 과제를 확인하고 마감일 전에 해결해 버립니다”라고 설명했습니다. 학생이 해야 할 일은 인기 있는 학습 관리 플랫폼인 Canvas에 대한 자신의 계정 정보(ID와 비밀번호)를 넘기는 것뿐이었으며, 아인슈타인이 나머지를 모두 알아서 해주겠다고 약속했습니다.
어떤 과제든 이 봇은 거뜬히 해냈습니다. 아인슈타인은 강의를 시청하고, 읽기 과제를 완료하고, 논문을 작성하고, 토론 포럼에 참여하며, 과제를 자동으로 제출할 수 있다고 자랑했습니다. 퀴즈나 기말고사가 온라인으로 치러진다면 아인슈타인은 기꺼이 그것도 대신 해결해 주었습니다.
처음 아인슈타인을 접했을 때 저는 회의적이었습니다. 화려한 AI 시연들은 과도한 약속을 하고 실망스러운 결과를 내는 경우가 많기 때문입니다. 그래서 직접 이 도구를 테스트해 보기로 했습니다. 대학생이 아닌 저는 무료 온라인 통계학 입문 강좌에 등록했습니다. 강좌 웹사이트는 이 수업이 자기 주도형이며 학부생, 대학원생, 의대생, 심지어 강사들에게도 기초적인 통계 지식을 쌓아줄 수 있다고 설명했습니다.
저는 이 봇을 풀어놓았고, 1시간도 채 되지 않아 아인슈타인은 8개의 모듈과 7개의 퀴즈를 모두 마쳤습니다. 약간의 걸림돌은 있었습니다. 봇이 한 퀴즈를 15번이나 다시 풀기도 했지만, 결국 그 수업에서 만점을 받았습니다. 저는요? 강좌 웹사이트를 읽어보지도 않았습니다.
관련 기사: AI 에이전트가 미국을 휩쓸고 있다
아인슈타인은 사람들을 자극하기 위해 설계되었습니다. 22살의 테크 기업가인 이 도구의 창작자 애드베이트 팔리왈(Advait Paliwal)은 교육자들에게 AI가 학교 과제를 얼마나 잘 해내는지 알리기 위한 방법으로 이 봇을 공개했다고 말했습니다. 그는 “날 탓할 수도 있겠죠. 하지만 이런 일은 지금 당장 일어나고 있고, 더 많은 사람들이 앞으로 다가올 일이 무엇인지 알아야 합니다”라고 말했습니다. (그는 이전에 AI에게 '사람들을 화나게 만들 웹사이트'를 만들어 달라고 프롬프트를 입력해 아인슈타인의 랜딩 페이지를 디자인했다고 밝힌 바 있습니다.)
아인슈타인을 공개한 직후, 팔리왈은 학업 부정행위를 영구화하도록 설계된 도구를 만들었다며 그를 질책하는 교수들의 이메일을 받기 시작했습니다. 그는 Canvas의 모회사 등 여러 곳에서 내용증명(Cease-and-desist) 편지를 받은 후 이 봇을 폐쇄했습니다.
팔리왈에게 이런 반발은 핵심을 벗어난 것이었습니다. 그는 “내가 이것에 대해 포스팅하지 않았다면, 누군가 같은 기술을 사용해 교수들에게 숨겼을 것입니다. 교수들이 이런 것이 존재한다는 것을 알고 앞으로 다가올 일에 올바르게 대비할 수 있다는 것이 사실 더 낫습니다”라고 말했습니다. 물론 이 도구는 팔리왈에게 바이럴한 명성을 안겨주기도 했습니다.
그럼에도 불구하고 아인슈타인은 교실에서 AI가 나아갈 방향을 보여주는 지표가 분명합니다. 최신 봇들은 거대한 컨텍스트 윈도우(Context windows)를 가지고 있어 학생들이 강의 계획서, 강의 슬라이드, 연습 문제와 같은 엄청난 양의 강의 콘텐츠를 입력할 수 있습니다. 오늘날의 에이전트 도구는 온라인 토론 포럼에 참여하고 녹화된 강의에 대해 메모를 작성하는 등 다양한 작업을 완료할 수 있습니다.
Listen − 1.0 x + Seek 0:00 9:19 William Liu is grateful that he finished high school when he did. If the latest AI tools had been around then, he told me, he might have been tempted to use them to do his homework. Liu, now a sophomore at Stanford, finished high school all the way back in 2024. “I have a younger sibling who is just graduating high school,” he said. “Our educational experience has been vastly different, even though we’re just two years apart.” By the time Liu graduated, ChatGPT was already causing chaos in the classroom. But the automation of school is intensifying. If at first teachers worried about students using chatbots to write essays, now new agentic tools such as Claude Code are allowing students to outsource even more of their work to the machines. Need to take an online math quiz? Write a biology-lab report? Create a PowerPoint presentation for history class? AI can do all of this and more. One high schooler recently told me that he struggles to think of a single assignment that AI wouldn’t be able to do for him. As a measure of just how good AI has become at schoolwork, consider a new bot called Einstein. Several weeks ago, the tool went viral with big claims: “Einstein checks for new assignments and knocks them out before the deadline,” a website advertising the bot explained. All that a student had to do was hand over their credentials for Canvas, the popular learning-management platform, and Einstein promised to do the rest. No matter the task, the bot was game: Einstein boasted that it could watch lectures, complete readings, write papers, participate in discussion forums, automatically submit homework assignments. If a quiz or a final exam was administered online, Einstein was happy to do that too. When I first came across Einstein, I was skeptical: Flashy AI demos have a way of overpromising and under-delivering. So I decided to test the tool out for myself. Because I’m not a college student, I enrolled in a free online introductory-statistics class. The course website explained that the class was self-paced and that it could help undergraduates, postgraduates, medical students, and even lecturers build up basic statistical knowledge. I set the bot loose, and in less than an hour, Einstein had worked through all eight modules and seven quizzes. There were some hiccups—the bot took one quiz 15 times—but it ultimately earned a perfect score in the class. As for me? I hardly so much as read the course website. Read: AI agents are taking America by storm Einstein was designed to provoke. Its creator, Advait Paliwal, a 22-year-old tech entrepreneur, told me that he’d released the bot as a way of alerting educators as to just how good AI is at schoolwork. “You can blame me,” he said. “But this is happening right now, and more people need to know about what’s to come.” (He has previously said that he designed Einstein’s landing page by prompting AI to make a website “that people would get angry over.”) Almost immediately after releasing Einstein, Paliwal started receiving emails from professors chastising him for creating a tool seemingly designed to perpetuate academic fraud. He took down the bot after he received multiple cease-and-desist letters, including one from Canvas’s parent company. To Paliwal, the backlash missed the point: “If I didn’t post about this, someone would have used the same technology and hidden it from the professors,” he said. “It’s actually better that they know that this exists, and they can correctly prepare for what’s to come.” The tool also, of course, gave Paliwal a moment of viral fame. Nevertheless, Einstein does seem to be an indicator of where AI in the classroom is headed. The latest bots have massive context windows , meaning that students can feed in mountains of course content such as syllabi, lecture slides, and practice exams. Today’s agentic tools can complete all kinds of tasks, such as participating in online discussion forums and taking notes on recorded lectures without student intervention. According to one analysis , the percentage of students middle-school age or older who self-reported using AI for help with homework climbed by 14 points from May to December of last year. Amid all of this, Silicon Valley is doubling down on its push to integrate AI into schools. In the lead-up to final exams last spring, nearly every major AI firm offered college students free (or heavily discounted) access to their paid chatbots. Now the tech industry is offering students cheap access to their agentic tools. Last summer, Anthropic announced “Claude Builder Clubs”—an initiative in which students paid by the AI company host workshops and hackathons on their campuses. In exchange for membership in those clubs, students are given free access to Claude Code. A few weeks ago, OpenAI announced that it would be offering college students $100 worth of credits for Codex, its agentic coding tool. The students affiliated with the AI companies, at least, say that the more powerful bots are helping them with their studies. Thor Warnken, an Anthropic ambassador and a biology major at the University of Florida, told me that he has designed what is effectively a personalized Khan Academy. When he takes a practice test—say, in organic chemistry—he feeds his completed work into Claude. He then asks the bot to find patterns in his errors and make new practice problems based on them. “The first practice question will be super easy, and the next one will get a little harder and a little harder, until it gets super hard,” he explained. Liu, who also serves as an ambassador for Anthropic, similarly said that the bot has made for a “fantastic” study partner. When he has questions during large lectures, he asks Claude, which has access to his course materials, and the bot explains concepts in real time; previously, those questions might have gone unanswered. Read: The AI takeover of education is just getting started Instructors, as I have previously written , are also using plenty of AI. Canvas recently introduced a new AI teaching agent designed to save instructors time on “low educational value tasks” such as organizing online-course modules and adjusting assignment due dates. “Faculty are using AI tools both for instructional purposes, for building course materials, but they’re also starting to play around with generative AI to actually grade and assess the learning,” Marc Watkins, a researcher at the University of Mississippi who studies AI and education, told me. He gave a hypothetical: “I could set my agent up, open it up in my course, go out on campus to walk across campus to get a cup of coffee at Starbucks,” he said. By the time he returned, 15 minutes later, all of the essays would be graded, and “bespoke personal feedback” would be sent out to each student. AI can save teachers time—that same grading takes him 10 or 12 hours, Watkins estimated—but in the process, the technology threatens the relationship between students and teachers that is core to education. “That’s really scary,” he said. Most people I spoke with seemed unhappy with the current trajectory of bots in the classroom. Even as growing numbers of students are using the technology, a majority believe that the more they use AI for classwork, the more it will harm their critical-thinking skills. Natalie Lahr, a Barnard sophomore studying history and political science, doesn’t use the technology “unless it’s something that’s asked of me by a professor,” she told me, “and even in that case, I’m generally quite opposed.” In one particularly “anti-AI radicalizing” experience, Lahr met with a tutor at the college’s writing center to get help on an essay. According to Lahr, that tutor copy-pasted her essay prompt into the popular AI tool Perplexity and gave Lahr the AI-generated outline. “That was basically the end of our session,” Lahr said. “I had a crashout about that afterwards because I was like, Why am I even here? ” Some educators are worried a