AI Guidance Report Updated

Do we say students can just learn biology on TikTok? If not, why do we leave them to learn about AI there?

Thank you again for supporting our work. If you find our guidance report helpful, please consider purchasing an annual subscription to this blog or picking up a copy of our book.

I updated our school AI guidance report to include ideas from the following AI guidance reports.

Michigan Virtual
National Education Association
North Dakota
Utah

This is a free document that you can download from SSRN.

Based on this work and a bit of other reading, I also want to share five quick thoughts.

Competition vs Co-Intelligence.  Many of the existing guidance documents still emphasize the importance of comparing human intelligence (HI) to AI, emphasizing the superiority of HI and arguing that humans must remain at the center of instruction and as the final decision-makers in education.  Based on the current state of the technology, this still makes sense, but it is important to acknowledge that AI is already competitive with HI in many areas and that AI exceeds it in some areas. For example, AI is better at humans at making predictions. Student placement and academic intervention, broadly speaking, are about predicting which interventions are appropriate in a given situation.  That’s what “personalized learning” is about.  When AIs can make better decisions because they can draw on more data, we are going to have to accept that humans will no longer be the best decision-makers in many instances.  And that’s why we need to see AI as “co-intelligence” or “augmented intelligence” –- as an extension of our intelligence.

Equality & equity. All of the documents stress the importance of equality and equity. What’s the difference? Equality in education aims to provide the same resources and opportunities to all students, equity focuses on ensuring that every student can reach and utilize those educational resources and opportunities, regardless of their individual circumstances or barriers. Lower SES students who attend public schools that restrict access to various AIS are less likely to have access (equality), but that gap in access is probably not really that large. Ninety-five percent of teens in the US have smartphone access, and students can access powerful Ais (ChatGPT4-0 and Claude Sonnet 3.5, for example) right from their smartphones. AI models are also now embedded locally in eyeglasses, and they will soon be embedded in contact lenses.

We do need to worry about the 5%, but equality in AI access is probably about 5% of the students. Equity is a far bigger challenge, with many students being taught how to use AI and others are left to on their own. To give a concrete example, thousands of students are learning with and about AI this summer in academic camps, including debate and science camps.  And the depth at which they are engaging the issues is far greater than anything they would ever do in school. These programs cost thousands of dollars, so many are excluded.

“What is Edtech?” Recent developments have started to get me thinking about what “edtech” is. This is a simple definition

Most edtech products (probably sans the internet and the calculator) have been provided to students through schools. But now that students can have free apps on their phones and computers that can teach them things, what uniquely defines “Edtech?” And given the restrictions schools have to put on these systems to comply with the law, might the “out-of-school ed tech” products that aren’t subject to those restrictions become more useful than the school-provided ones?

AI Literacy. All the guidance documents point to the importance of developing AI literacy. Like Vera Cubero, if I had the power, I’d make every student take an AI literacy class next year, and I’ve worked on some projects that aimed to develop AI literacy curriculums.

But I’m a dreamer. It’s just not realistic that every school has the resources (money, personnel) to implement such a course. It’s not clear where it would go in the schedule. Is it practical to just add stuff? Where is the fully structured and scaffolded curriculum? What world are educators willing to accept they need to prepare students for? Many do not see how radically AI is starting to alter the world.

Implementing a full-blown AI literacy course is beyond what most schools can do, but, as I’ve mentioned before, students can learn about AI through debates, AI-related lessons in particular courses, and cultivated AI news stories set for discussion (more on that later).

We do need to work on finding a way to teach students with and about AI. If we don’t, they’ll just learn about it where they do now — on TikTok. Do we say students can just learn biology on TikTok? If not, why do we leave them to learn about AI there?

Professional development. Without a solid understanding of AI, teachers cannot effectively comply with school AI policies, support their students' use of AI tools, or adjust their assignments to account for AI capabilities. As AI becomes more prevalent in education and society at large, teachers need to be equipped with the knowledge to guide students in using these technologies ethically and productively. Moreover, understanding AI allows educators to redesign their curricula and assessment methods to foster skills that complement, rather than compete with, AI capabilities. This knowledge empowers teachers to prepare students for a future where AI literacy will be as fundamental as digital literacy is today.

I strongly believe that AI PD needs to be honest. AI is a very powerful technology, and that power must be shared with teachers. AI is a tool that can learn largely on its own, share knowledge in an instant, share more knowledge than any human, do more and more tasks and jobs, and reduce the value of some skills (language translation, for example). It’s not another piece of Edtech.

Rate of change. The pace of AI development is currently exponential, a rate of progress that is notoriously difficult for the human mind to grasp or predict. This rapid advancement is challenging our understanding of technological growth and its societal impacts. The concept of exponential progress means that the capabilities of AI systems are not just improving gradually, but doubling at regular intervals, leading to sudden and dramatic leaps forward. This phenomenon is evident in various domains: AI systems now fly planes without human pilots, demonstrating complex decision-making in three-dimensional space, while Waymo's autonomous vehicles have safely navigated over 10 million miles without human intervention. These achievements, unimaginable just a few years ago, illustrate the accelerating curve of AI advancement.

This reality creates challenges for education, but it’s still a challenge we need to respond to.

Metacognitive development. Metacognitive skills refer to the ability to think about one's own thinking processes. These skills include self-awareness, self-regulation, and the capacity to plan, monitor, and evaluate one's learning and problem-solving strategies. In an AI-driven world, metacognitive skills are increasingly crucial for students. As AI systems become more adept at information retrieval and routine problem-solving, human value increasingly lies in higher-order thinking, creativity, and the ability to navigate complex, ambiguous situations.

Students with strong metacognitive skills can better discern when and how to use AI tools effectively, critically evaluate AI-generated information, and adapt their learning strategies in response to rapidly changing technological landscapes. These skills enable students to become lifelong learners, capable of adjusting to new challenges and technologies throughout their careers.

Schools can foster metacognitive skill development through various approaches, even without specific knowledge of AI:

Project-based learning: By engaging students in long-term, complex projects, schools encourage planning, self-monitoring, and reflection. Students learn to break down large tasks, manage time, and adjust strategies based on progress.

Debates and discussions: These activities promote critical thinking, perspective-taking, and the ability to construct and evaluate arguments. Students learn to question assumptions and consider multiple viewpoints.

Reflective journaling: Regular reflection on learning experiences helps students develop self-awareness and the ability to analyze their own thought processes.

Problem-solving activities: Presenting students with open-ended problems encourages them to develop and evaluate multiple solution strategies.

Peer and self-assessment: These practices help students develop the ability to critically evaluate work, including their own, fostering self-awareness and analytical skills.

Mindfulness and metacognitive exercises: Explicit instruction in metacognitive strategies can help students become more aware of their thinking processes.

Collaborative learning: Group projects encourage students to verbalize their thinking, consider others' perspectives, and collectively plan and evaluate strategies.

All of the documents stress the importance of these skills, but most give a short rift to implementing them. Given how rapidly AI is advancing and the importance of metacognitive development in adapting to this world, we strongly encourage schools to act soon to support these programs.