AI in Education: Reshaping Education for the First Time in Over 200 Years
Our education system desperately needs a major overhaul. Today’s system costs way too much, is unavailable to some who need it and doesn’t teach students the skills they will need to succeed in the future workplace. AI in education is about to change that.
Academia is notoriously resistant to change. And with good reason—they don’t want flavor-of-the-month experiments conducted on their students. But in the age of AI, our education system must stop resisting and start embracing a transformative technology that will enable it to scale its efforts, deliver better educational outcomes, significantly cut costs, and make education accessible to all. Saddling students with a lifetime of debt is not a great way to build a vibrant economy and society.
AI is not a threat to educators. Not if they embrace it. Artificial intelligence offers powerful tools and capabilities to help educators do what they do, but way more effectively. AI will improve every aspect of education, from marking exams to building course materials, optimizing curriculums, and augmenting teaching with personalized tutors that allow students to learn at their own pace and learning style.
First, Let’s Stop Vilifying ChatGPT
The number one concern I hear from educators is some version of: “But how do we stop students using ChatGPT to cheat?” This is like banning calculators from a math exam. Mental arithmetic skills are essential, but in the real world, mathematicians and engineers have full access to calculators and computers and use them to make calculations and solve problems. AI tools are never going away. Artificial intelligence is the worst it will ever be and will improve exponentially in the coming years. AI will be part of our daily lives and deployed throughout the workplace. Educators must prepare students to perform in a world of AI-powered tools, AI assistants, and machine coworkers. They should teach students how to achieve optimal results working with AI, not without it. Historically, the best way we knew to measure knowledge and a student’s ability to wield it as part of a cogent set of arguments was to force them to write essays that students don’t want to write and that educators don’t particularly enjoy reading and assessing. Essay writing has essentially become ‘commoditized’ by large language models like ChatGPT, Gemini, and Claude. We need to find new ways to assess students. And AI can help.
AI in Education
There are four main areas where artificial intelligence will transform the future of education:
Curriculum - Preparing students for the future world of work where AI is everywhere
Pedagogy - Changing how we teach students using AI
Research - Transforming and accelerating academic research with powerful AI tools and assistants
AI talent pipeline - Expanding courses to train the talent needed to invent, build, and deploy AI
Let’s examine each individually and explore how AI presents significant opportunities to transform education, improving the experience for tutors, administrators, and students alike.
Curriculum for the 21st Century
The world of work is transforming rapidly. Widespread deployment of AI will utterly transform work in the coming decade. We must prepare students to thrive in a workplace that doesn’t yet exist, where coworkers are machines and AI assistants augment human capabilities and elevate and amplify employees’ skills. ChatGPT is just the start of the AI-based technology that will be available to workers in the near future. Existing workplace tools—Microsoft Office, Google Workplace, the Adobe Creative Suite, etc.—are turbo-charged with AI. New tools emerge daily that harness AI to boost employee productivity and increase the quality of their output. Students must learn how to get the most out of these tools and how to work alongside AI. They will work in a world of autonomous AI assistants (essentially a fellow employee built from software) and be responsible for constantly reengineering their workflows as new AI capabilities emerge, using process orchestration and AI transformation to offload lower-value tasks and focus on higher-value tasks that require human attention.
Open questions:
How do we prepare everyone for the future of work?
What new skills will people need to survive and thrive in the workplace of the future?
What skills are we currently teaching that will be less relevant and might be deemphasized?
How could we scale the education system so academia can help more students AND reskill/upskill adults already in the workforce? (A HUGE opportunity for academia!)
Pedagogy for the Intelligence Age
Education hasn’t changed much since Victorian times when this image was taken
Today’s education system has changed very little from the one built in the Industrial Revolution to prepare farm workers for careers in factories and offices. Whiteboards and PowerPoint presentations have replaced chalkboards, tablets and laptops have replaced notepads, and lab equipment is more sophisticated. Still, the prevailing pedagogy is a teacher standing at the front of a classroom lecturing students, who furiously take notes and later are assigned homework and project work.
If we were designing an educational system from scratch today that embraced all the high-tech available, we certainly wouldn’t design it the way it is. As AI is widely adopted in the workplace, workers displaced by inevitable productivity gains and automation will need support from the education system to reskill, upskill, and maintain their relevance and marketability. More people will need access to high-quality, low-cost education than ever before. The education sector must step up to the plate to modernize our economy and help our population escape a rising tide of automation.
Today’s education model is inherently difficult to scale, limits the progress of fast learners, and risks leaving other students behind. An AI tutor won’t displace human faculty but offers a compliment to their teaching endeavors. An AI tutor, well implemented, will help students learn with the Socratic method (helping them to ask the right questions but never providing the answer). Each student will learn at their own pace. AI tutors will adapt to each student’s learning style and spend as much time with students as they need.
A great early example of an AI tutor is Khanmigo, built by the Khan Academy. The tutor, built in partnership with Microsoft and based on OpenAI’s latest technology, GPT-4o, is an impressive service available to students and parents for just $4/month. The Khan Academy also offers a free AI teaching assistant for teachers, helping them with lesson planning and creating discussion prompts, multiple-choice assessments, quiz questions, rubrics, and more. This video shows Khan Academy founder Sal Khan using Khanamigo to help his son learn simple trigonometry.
Many AI-powered tools have already been developed for the education sector, primarily aimed at faculty and administrators. Here is a small sample of what’s available:
Education Copilot - Generate lesson plans, handouts, and other education materials in seconds.
ClassPoint - Easy generation of quizzes, presentations, word clouds, and more.
Magic School - Syllabus generator, teacher assessment, lesson planning, and more.
Gradescope - A grading tool to assess science, math, and economics assignments, exams, homework, and even software code.
Additio - Automatically create schedules and timetables to optimize class assignments, faculty, and spaces.
AI Media - Real-time language translation and classroom captioning to level the playing field for all students.
AvidNote and Kahubi - Read and write research papers faster and analyze your data with AI tools.
ResearchRabbit - Research assistant. Free discovery, analysis, and connection tools (including to funders!)
Consensus - AI-powered search engine for research.
Many more tools are in development, so look for new AI-powered tools that solve your particular educational challenges.
An Important Warning
The Khan Academy’s mission is to provide a free, world-class education to anyone, anywhere. People working in the traditional education sector should pay close attention to this statement, internalize it, and understand the scale of the existential risk it poses, perhaps not from Khan Academy, but from others. If academia fails to embrace AI, improve their offering, and significantly reduce the cost of high-quality education, they risk disruption by well-funded, highly motivated external competition. Amazon Prime Education? Microsoft AI Tutor? Google University? All it will take is for high-tech companies to convince employers that their accreditation is as good a way (or better?) to assess the skills and knowledge of prospective job candidates for the future workplace. When that happens, the high price tags of traditional education will become hard to justify. Beautiful university campuses may become the abandoned shopping malls of the 2030s, redeveloped into hotels, offices, and leisure facilities, or abandoned to the elements. This is not hyperbole. Education will change whether the calcified world of academia chooses to welcome change or not. It’s a simple choice: Embrace AI and rethink pedagogy for the 21st century to improve outcomes and slash costs or become irrelevant. Community colleges will go first, then universities (including Ivy League institutions) will follow.
Open Questions
How could AI be used to improve access to education and significantly reduce the cost of delivering a high-quality education to all?
What is the right mix of in-person human tuition versus self-guided, self-paced learning with an AI tutor? How might that vary from person to person? How can AI tutors coach students during homework and assignments to solidify their learning?
How can academic staff use AI to make their lives easier and scale their impact and productivity 10x?
Accelerate Research to Push the Boundaries of Human Knowledge
As tools like Google DeepMind’s AlphaFold have demonstrated, AI turbocharges scientific and medical research. Discovering the structure of a single protein required five years of graft by a PhD biology PhD. With AlphaFold, that time was reduced to seconds. AlphaFold unlocked the structures of all 200 million known proteins and opened the door to research on pharmaceuticals and other therapeutics that will disrupt disease paths and potentially yield cures for Alzheimer’s, Parkinson’s, and cancer. Those same PhDs can now focus on unlocking new problems, not the tedious work of crystallography to determine protein structure.
AI will quickly become an invaluable tool for all researchers, expanding their thinking and helping them to explore the boundaries of scientific understanding.
Here are examples of some early AI tools available to researchers:
ResearchRabbit - Research assistant. Free discovery, analysis, and connection tools (including to funders!)
Consensus - AI-powered search engine for research.
Machines will be partners with human researchers across every scientific discipline. They will help us think and provide a new canvas for scientists to explore the universe's secrets. In the same way that we externalize and aid thinking through symbols and diagrams on a blackboard, AI will support a new level of thought elevation by helping us capture and build upon our ideas. AI will likely transform research even more than educational pedagogy and curriculums.
Open questions:
How might research be conducted differently in the future if every researcher had access to AI assistants trained in their field of research?
How could academia develop and embrace AI tools that accelerate research and significantly boost global impact?
Building the AI Talent Pipeline
During my talks on AI, audiences often ask me what jobs AI will create. New technology always destroys some jobs while creating others. The railways made life hard for horse breeders, cartwrights, and blacksmiths but created many new roles: rail engineer, train driver, stationmaster, timetable maker, buffet car server, signalman, maintenance engineer, and many more.
In the AI era, we will need people trained to invent, build, and deploy AI technology at scale—people with training in machine learning, data science, data center operations, AI chip design, cybersecurity, and more. We will need systems architects, AI solutions designers, AI application coders (who know how to get the most out of AI coding assistants), prompt engineers, model tuners, data center designers, process orchestrators, AI sales and marketing experts, and many more roles that have yet to be invented. AI talent is in high demand. Top AI researchers can command seven-figure salaries.
We are beginning a multi-decade rollout of AI technology across our society. Every company will need a Chief AI officer; every IT department will need to fill its ranks with AI expertise, and the data centers that power AI will all need to be built, operated, and managed by someone. A new data center is currently built somewhere in the world every three days. As we build out the AI capacity needed to turbocharge business, education, healthcare, and government operations, some estimate we will need 100 times the number of data centers that exist today.
In Summary
While they may not yet acknowledge it, educators and every other industrial sector are about to be transformed by the coming of AI. Not changed a little bit. Not doing the same things as before but allowing ChatGPT into the classroom. The wholesale and much-needed transformation will make high-quality education accessible to many and complement human capital with AI smarts that enable us to scale education while significantly reducing costs. A tsunami of change is coming. Leaders in higher ed, and K-12 for that matter, can either figure out how to surf that wave or risk being wiped out by it.
If you’re interested in learning more about AI's transformation of education, hire me to deliver a keynote to your team. My latest keynote, Revolution Ed: The Urgent Need to Reinvent Education in a World of AI, explains what AI can do, where it’s going, and what you can do to begin your AI transformation journey today. Details on my website are here.