AI’s Quiet Tsunami #4: Next, AI Resets Education and Learning

(I hope you've read the other three in this series: AI’s Quiet Tsunami , Challenges in a World Fueled by AI , and AI and (Your) Work .)

So where do you stand?

• The new crop of AI tools is a plague on education . (And educators aren’t prepared.) Or…
• The new tools are the reboot that education needs, opening the door for a new paradigm of learning. (And educators still aren’t prepared.)

No matter where you stand, this is a watershed moment. Here are four takeaways:

• This is an Internet Moment for education. The global pandemic served as training wheels for leveraging technology to enable new ways of learning, which in hindsight will serve as preparation for the new AI tools.?
• AI tools are only a threat if we keep using “the old rules of education.”
• The new AI tools offer a tremendous opportunity to transform the process of learning.
• To do this, we must reimagine many of the fundamentals of education.

This is an Internet Moment for Educators

In 2010, our managing partner Heidi Kleinmaus and I co-produced a conference on the future of Higher Education. (My favorite attendee tweet: “The startups are pitching, and the academics are bitching.” Innovators saw only opportunity: Incumbents saw only threat.) In 2014 I wrote Unbundling Higher Education , suggesting a paradigm that leaves behind our industrial-era model of institutionalized learning. But change in our institutions of learning is historically glacial.

Then, along came a virus.

The global pandemic was a great reset for education, requiring overnight adaptation by educators and students around the world. It illustrated how rapidly we can change, given the “right” combination of incentives and disincentives. It also showed clearly how we have trained today’s teachers in an industrial-era model that makes no sense in a world of dramatic change.?

But I have tremendous compassion for teachers and students in this era. Being part of such dramatic change is hard. The calculator, the PC, the Internet, the global pandemic… Each required teachers to adapt, with little preparation or training. Some people say change is inevitable, technological progress is inexorable, blah blah blah. There is an unquestionable human toll from exponential change, and unless we train teachers and provide the appropriate tools, the job of an educator will simply become so challenging that far fewer people will choose the profession.

Yet it’s completely understandable if, today, many see AI’s quiet tsunami as a challenge to much of what they’ve been trained for.

AI as Threat

• “What winter 2020 was for Covid, winter 2023 is for ChatGPT…A return to handwritten and oral in-class assignments — a lockdown response — may be the only immediate effective solution as we wait for more robust protections to arise.” — Opinion piece in Inside Higher Education

If you are in the “threat” camp, these tools are a direct assault on the ways traditional education has been taught for a century. ChatGPT is the greatest cheating tool ever invented . Since students are already using these tools , it’s the end of High School English , or even writing as we know it . It could even mean the end of grading . Instructors and authors are writing 2,000-word academic papers in 10 minutes , writing common app essays that fool admissions officers , and creating syllabuses (syllabi?) for MBA-level courses ,?

Of course, not all instructors are freaking out , since tools like ChatGPT-3 aren’t yet great, for example, at 12th-grade AP literature . Some instructors are running their tests through ChatGPT, and dismissing the software because it only got a B . That’s like saying a horse can beat a Model T, when a Ferrari is just around the corner. The tools will get better, fast.

Right now, for 4th and 8th-grade creative writing — they’re not bad . And before long, AI software will be able to pass exams from law and business school , or even get an MBA ,?

There are already many tools to detect plagiarism by students, and for deepfake videos . Many teachers are calling for similar detection tools for AI . OpenAI researchers announced a method called GPTzero that could detect about 20% or more of generated text, depending on length. (More text, better results .) In a little over a month, 43,000 teachers signed up.?

That’s an understandable reaction. But it’s like asking students to stay on a horse while others are learning to drive cars: students leveraging the new AI tools will have greater fuel for their learning. And one educator said that it’s ridiculous to try to stop cheating . It’s not just trying to close the barn door after the horse is gone. The barn is on fire.

What the new AI tools can — and should — do is to encourage us to completely transform the way we teach .

AI as Opportunity

There are already many non-AI tools to help teachers in every facet of education. The new AI tools can serve as a huge leap forward. Teachers who are instead embracing these tools are finding a challenging but exhilarating new world of learning. Just a few opportunities include:

• Generating “first draft” content. Essay writing can and should be completely rethought . Some teachers are even requiring students to use the new tools.
• Fueling rapid learning. Research shows that the pace of time-intensive learning like phonics can be dramatically increased with automation.
• Making education fairer . It’s very likely that tools like ChatGPT will help struggling students the most .
• The digital teacher’s aide. AI software can generate multi-choice questions , grade homework , and design authentic assessments ,?
• The administrator’s aide. AI tools can automate record-keeping and data management ,?

Of?course, the new tools won’t serve as magic wands to transform education. Teachers have heard that song many times before. Instead, we need to see this as an opportunity to transform formal education itself, so we can leverage the new tools more effectively.

We Must Reimagine Education

I want to stretch your mind a bit. So here’s a question for you: What’s the difference between education and learning?

You might say that education is the set of institutions that provide formal learning. Or you might say that Education is a noun, and we all need one. (AI-fueled writing structure would urge me to paste in ChatGPT’s answer to the question, since many humans seem to treat the software like a modern Oracle . Let’s not.)

But what if we simply thought of Learning? How might we act differently? If we conceived of the ideal process by which all little bitty humans could become high-function big humans, what would that look like?

Well, we know that there are predictable phases that each of us goes through, developing our minds, our bodies, our emotions, and the skills for interacting with others and the world around us. We will have predictable roles in life, such as child, sibling, parent, worker, community member, and citizen. And there are several predictable goals that each of us will have, which can include any or all of having fun and being happy, finding a mate, having children, gaining knowledge, and finding or creating meaningful, well-paid work.

We also know that different people are drawn to developing different beliefs, values, and skills, through a complex Venn diagram of their genetic characteristics, parents and other family, community, culture, and experiences.

So, given all this variety of human experience, how could we design our systems of learning?

Research is quite clear how the typical human learns best. Trial-and-error. Play-based learning when we’re smallest, project-based learning as we age. Interest- and passion-based learning. Solving real-world problems. Collaborative problem-solving. Help getting through learning speed bumps. Learning from and with other children, older and younger. Some, but not too much, repetition. Immediate application of learning to action.

Now, let’s look at how traditional Western schooling fails to incorporate many of these insights.

• Only about half of U.S. kids have some kind of learning environment before kindergarten. So we’re failing the other half of them, right out of the gate.
• In kindergarten, a growing movement of concerned parents push for “academic” learning, so their children won’t “fall behind.”
• We lock children into cohorts of the same age and both sexes. (We call these “Grades.”)
• We tell them that learning happens in one physical location. (“School.”)
• They are immersed in rote learning rarely connected to any real-world use. (“Drills.”)
• We don’t let them do all their learning in a context with others: We require them to do an increasing amount of dubious practice at home, alone. (The “Homework ” Machine.)
• Frequent evaluation of rote memorization. (“Tests.”)
• Stigma for giving the “wrong” answer, rather than for demonstrating a creative and/or effective thought process. (“Failing.”)
• One-size-fits-all learning, with little adaptation for learning differences, prior trauma, or limited outside support. (“Common Core Curriculum.”)
• Stage-gates to move to the next phase of the cohort. (“Graduation.”)
• One-size-fits-all recognition of passing through a major stage-gate. (“Diploma.”)
• Social and emotional learning as a possible but not guaranteed by-product, rather than as a set of skills to intentionally develop.

As longtime tech journalist and Arizona State University professor Dan Gillmor says , “Academia has some serious issues to confront,”

Now, how might artificial intelligence and other exponential tools change the game?

• Creativity. Using AI tools to generate “first draft” content can help to open up the range of creative options they could explore.
• Insights: Software can help us understand the amount of knowledge a learner has, and the ways they learn best.
• Repetition: Software is infinitely patient, when repetition is actually useful for certain kinds of learning.
• Critical thinking . Have students generate three different texts or images, then compare and synthesize them. Where are the tools “right”? Where are they wrong ? How do you know?
• Customized learning: AI can use questions and quizzes to understand how a learner thinks, and what they already know. For an example, check out Squirrel AI Learning in China, which does exactly that to customize math and science in K12. (Check out this panel I moderated with Squirrel co-founder Derek Haoyang Li.)
• Rethink Homework : Focus on accelerated in-classroom learning with other learners.
• Remembering: Software can retain its understanding of the learner , knowledge which is often lost in traditional education in the passage from one teacher to the next.
• Collaborating: As young learners develop their own superpowers, they can better understand the strengths of others to enhance their own strengths.
• Networking: Research shows that the social connections between students can nearly be as powerful as the education itself. Especially in high schools in urban areas, AI can help guide students to diverse schools and after-school programs where they can build diverse networks.
• Encouraging exploration: Graded homework isn’t an effective way to teach most subjects. Instead, students following their own passions and interests should do so without worrying about (often-arbitrary) grading.

The net Next: Teachers will need to rethink the very basis of student learning — in collaboration with students. We will all have to co-create the rules for the kinds of learning that need to take place in the classroom and beyond. We have inherited the “old rules” of learning from the industrial era. If you think of the great reset of the pandemic as a set of training wheels for change, the new AI tools mean we need a completely new bicycle. Say hello to personalized and self-paced learning, homework rooted in critical thinking and collaborative problem-solving over rote memorization, and creative demonstrations of learned skills.

For more ideas, check out this recent talk I delivered in Medellin, Colombia, A Vision for an Exponential Future of Learning , at the Virtual Educa conference.

What Can We Do, Next?

• Learners: We must teach students to become intelligent consumers of all media, especially the new AI tool. That includes understanding when software starts to sound like people .
• Teachers: Start with your own learning and experimenting . Mobilize learning cohorts and crowdsource ideas to help teachers understand and leverage the new tools. Learn best practices from each other, just as you did during the pandemic. But we really need to teach teachers to teach differently : Old methods of simply penalizing for the use of assistive technologies won’t fly any more. (Read Esther Wojcicki 's Moonshots in Education . And check out Marc Prensky — EMPOWERING ALL YOUNG PEOPLE 's Education to Better Their World .)
• Administrators: Treat this with the urgency it requires. Start an internal task force in your school or university. (I call these NEXTLabs.) Reward teachers who embrace the new tools, and who help other teachers to use them.
• Parents: Practice with these tools yourself. You don’t have to be an expert. But you’ll be able to discuss their use with your kids.
• All of us: Demand that software developers create tools to help educators and students to use them to improve learning. This is a fast-moving train: The pace of new tools, and revisions to existing ones, will come fast and furious. The toolmakers need to focus especially on their use in learning environments, and make them deeply accessible.

-gB Gary A. Bolles

• I’m the author of?The Next Rules of Work: The mindset, skillset, and toolset to lead your organization through uncertainty ;?the adjunct Chair for the Future of Work for?Singularity Group ;?have over 1.3 million learners for my courses on?LinkedIn Learning ;?a partner in the consulting firm?Charrette LLC ;?a co-founder of?eParachute.com ; an original founder of?SoCap Global , and former editorial director of 6 tech magazines. Learn more at?gbolles.com