[CONFERENCE] 'Untutored AI will be disaster,' KCL president says

King’s College London Vice-Chancellor and President Shitij Kapur said universities must teach people to work with AI and not for it in an online keynote at the 2026 Korea Times Global Conference in Seoul on Tuesday. He called for a rethink of learning, assessment and graduate preparation as AI reshapes education, work and research. Kapur said universities should use scaffolded AI, protect slow creative scholarship and focus more on experiential, emotional and relational learning. He also warned against mistaking AI’s fluent imitation of empathy for real human intimacy.

Kapur said universities have five central responsibilities in the AI era: redefining learning, preparing graduates for a changing labor market, defending slow scholarship, promoting AI for the common good and ensuring AI enhances humanity.

He argued that AI is pressuring traditional university models built on lectures, essays and degrees, and said untutored AI would be a disaster for the university.

He said randomized controlled studies suggest AI can improve immediate output without necessarily deepening learning, while scaffolded AI can improve student performance, especially for lower-performing students.

Kapur said universities should shift more educational focus from cognitive and analytical work toward experiential, emotional and relational learning.

He cited a recent OpenAI effort involving the Navier-Stokes equations, saying humans had not solved the problem across the world for 70 years before it was solved in 88 hours using 10,000 agents, 3 million messages and $15 million.

Published Sep 15, 2026 3:20 pm KST

Updated Sep 15, 2026 5:08 pm KST

Kapur calls for more experiential education, protection of creative human work

King’s College London Vice-Chancellor and President Shitij Kapur delivers a keynote address online from London during the 2026 Korea Times Global Conference at the Korea Chamber of Commerce and Industry building in Seoul, Tuesday. Korea Times photo by Shim Hyun-chul

Universities must ensure that people work with artificial intelligence (AI) and not for it, King’s College London Vice-Chancellor and President Shitij Kapur said, calling for a fundamental rethink of learning, employment preparation and the public purpose of higher education in the age of AI.

Delivering a keynote speech at the 2026 Korea Times Global Conference, titled “Beyond AI: Building human-centered leadership for next era,” Kapur said universities face a defining task: shaping how AI affects education, work, research and society as a whole, rather than simply adapting to the technology’s commercial imperatives.

“The slogan for universities, and dare I say, for humanity has to be that we will work with AI and not for AI,” he said in his online address from London on Tuesday.

He argued that AI will not produce a single universal future, as countries will develop and apply AI according to their own needs and ambitions. Still, he said, universities everywhere have five central responsibilities: redefining the essence of learning, preparing graduates for a changing labor market, defending slow and creative scholarship, promoting AI for the common good and ensuring that AI enhances humanity.

Kapur said AI has put traditional university education under pressure, challenging lectures as the main method of teaching, essays as the principal form of assessment and degrees as the core credential.

According to him, randomized controlled research studies suggest that AI can improve immediate output without necessarily deepening learning.

“So untutored AI will be a disaster for the university,” Kapur said.

Rather than banning AI or allowing unrestricted use, universities should deploy “scaffolded AI,” he said — tools designed around a course’s learning objectives and used with guidance. Such systems can improve student academic performance, with the greatest gains among lower-performing ones.

Kapur also said university education will need to place less emphasis on purely cognitive and analytical tasks as AI takes over more reading, writing and routine knowledge work. Schools should therefore shift more of their educational focus from cognitive and analytical work toward experiential, emotional and relational learning, he added.

In this sense, universities must also preserve their function as communities where students develop personal and professional identities, Kapur said. That role will become more critical today as AI reshapes not just how students learn, but also how they work after graduation.

Kapur said universities should also resist equating speed and scale with intellectual value. AI may be able to process huge volumes of information and solve some research problems like the Navier-Stokes equations faster than any humans can. But higher education must keep in mind that it has a responsibility to protect areas in which slow, creative and scholarly work remains essential, he said.

“So here is a problem humans were not able to solve across the world, talking to each other, for 70 years. And here is OpenAI, 10,000 agents, 3 million messages, $15 million, solved in 88 hours,” he said, citing what he described as a recent OpenAI effort involving the Navier-Stokes equations. “We should actually try and identify the places where slow, creative scholarly work is essential, and then defend those places.”

Kapur concluded that the crucial task is to protect human agency, genuine relationships and the embodied, finite nature of human life. AI may communicate fluently and even imitate empathy, but people should not mistake such simulations for genuine human intimacy. Nor should societies inadvertently confer legal agency on AI systems, he added.

“To be human is not merely a thing,” he said. “It is to author one’s life.”

He said the role of universities is not only to develop and adopt the latest technology, but also to remind society of what should matter in an AI-driven world.

伦敦国王学院副校长兼院长希蒂吉·卡普尔（Shitij Kapur）周二在首尔举行的2026年韩国时报全球大会上发表线上主题演讲时表示，大学必须教会人们如何与人工智能合作，而不是为人工智能服务。他呼吁重新思考学习、评估和毕业生培养模式，因为人工智能正在重塑教育、工作和研究。卡普尔表示，大学应该使用循序渐进的人工智能，保护缓慢而富有创造性的学术研究，并更加注重体验式、情感式和关系式学习。他还警告人们不要将人工智能流畅地模仿同理心误认为是真正的人际亲密关系。

卡普尔表示，在人工智能时代，大学有五项核心责任：重新定义学习，让毕业生为不断变化的劳动力市场做好准备，捍卫缓慢的学术研究，促进人工智能造福大众，并确保人工智能提升人类福祉。

他认为人工智能正在给以讲课、论文和学位为基础的传统大学模式带来压力，并表示未经训练的人工智能对大学来说将是一场灾难。

他表示，随机对照研究表明，人工智能可以提高即时产出，但不一定能加深学习，而支架式人工智能可以提高学生的表现，尤其是对表现较差的学生而言。

卡普尔表示，大学应该将教育重点从认知和分析性工作转向体验式、情感式和关系式学习。

他引用了 OpenAI 最近一项关于纳维-斯托克斯方程的研究，称人类在全世界范围内花了 70 年时间才解决这个问题，而 OpenAI 仅用了 88 小时，就利用 10,000 个智能体、300 万条消息和 1500 万美元解决了这个问题。

发布于2026年9月15日下午3:20（韩国标准时间）

更新于2026年9月15日下午5:08（韩国标准时间）

卡普尔呼吁加强体验式教育，保护人类的创造性工作。

周二，伦敦国王学院副校长兼院长希蒂吉·卡普尔（Shitij Kapur）在首尔韩国工商会大楼举行的2026年《韩国时报》全球大会上，通过网络发表了主旨演讲。照片由《韩国时报》沈贤哲（Shim Hyun-chul）拍摄。

伦敦国王学院副校长兼院长希蒂吉·卡普尔表示，大学必须确保人们与人工智能 (AI) 合作，而不是为人工智能工作。他呼吁在人工智能时代对学习、就业准备和高等教育的公共目标进行根本性的反思。

卡普尔在 2026 年韩国时报全球大会上发表了题为“超越人工智能：为下一个时代构建以人为本的领导力”的主题演讲。他表示，大学面临着一项决定性的任务：塑造人工智能如何影响教育、工作、研究和整个社会，而不是仅仅适应这项技术的商业需求。

“大学的口号，我甚至可以说全人类的口号，必须是与人工智能合作，而不是为人工智能服务，”他在周二于伦敦发表的在线讲话中说道。

他认为，人工智能不会带来单一的普适未来，因为各国会根据自身的需求和目标发展和应用人工智能。但他同时指出，世界各地的大学都肩负着五项核心责任：重新定义学习的本质，使毕业生适应不断变化的劳动力市场，捍卫稳健而富有创造性的学术研究，促进人工智能造福全人类，并确保人工智能能够提升人类福祉。

卡普尔表示，人工智能给传统的大学教育带来了压力，挑战了以讲授为主要教学方法、以论文为主要评估形式以及以学位为核心资格认证的方式。

他表示，随机对照研究表明，人工智能可以在不加深学习的情况下提高即时输出。

“因此，未经训练的人工智能对大学来说将是一场灾难，”卡普尔说。

他表示，大学与其禁止人工智能或允许其无限制使用，不如部署“支架式人工智能”——即围绕课程学习目标设计并在指导下使用的工具。这类系统可以提高学生的学业成绩，尤其对成绩较差的学生提升效果最为显著。

卡普尔还表示，随着人工智能承担越来越多的阅读、写作和常规知识工作，大学教育需要减少对纯粹认知和分析任务的重视。他补充说，因此，学校应该将教育重点从认知和分析工作转向体验式、情感式和关系式学习。

卡普尔表示，从这个意义上讲，大学也必须保持其作为学生发展个人和职业认同的社区的功能。随着人工智能不仅重塑学生的学习方式，也重塑他们毕业后的工作方式，这一作用在今天将变得更加重要。

卡普尔表示，大学也应该避免将速度和规模与智力价值划等号。人工智能或许能够处理海量信息，并比人类更快地解决一些研究难题，例如纳维-斯托克斯方程。但他指出，高等教育必须牢记，自身有责任保护那些需要缓慢、创造性和学术性工作才能完成的领域。

“所以，这是一个人类在全世界范围内交流了70年都无法解决的问题。而OpenAI，1万个智能体，300万条信息，耗资1500万美元，只用了88小时就解决了这个问题，”他说道，并引用了他所说的OpenAI最近在纳维-斯托克斯方程方面的一个项目。“我们应该努力找出哪些领域需要缓慢而富有创造性的学术研究，并捍卫这些领域。”

卡普尔总结道，至关重要的是保护人类的自主性、真挚的人际关系以及人类生命的具身性和有限性。人工智能或许能够流利地交流，甚至模仿同理心，但人们不应将此类模拟误认为真正的人际亲密关系。他还补充说，社会也不应在无意中赋予人工智能系统法律代理权。

“为人不仅仅是一种身份，”他说，“而是要书写自己的人生。”

他表示，大学的作用不仅在于开发和采用最新技术，还在于提醒社会在人工智能驱动的世界中什么才是最重要的。