Category: AI & Education

When a technology offers advantages and disadvantages, the decision-making process can be quite complicated, especially when oversight cannot be guaranteed. For example, many states now ban cell phones, making a use such as telling parents when the schedule for after-school activities has changed difficult. The advantages and disadvantages vary with the field of application and my interests have mainly been focused on education. Just to be clear, by this I mean learning in general, not just the type of learning that occurs under supervision or is associated with educational institutions.

The generic educational situation that raises concern involves tasks undertaken to encourage both skill development and knowledge, and includes a requirement that demonstrates that the task has been attempted by the existence of some product. In educational settings, such products might result from homework or class activities, or simply by visible demonstrations of activity. The issue with AI is that in many cases, such as problem sets or documents of various types, these same products could be generated by AI, avoiding the cognitive activity of the learners. The phrase “cognitive offloading” has been used to describe this alternative form of product creation. Teachers might simply call it cheating. Cognitive offloading itself can be a desirable or undesirable option, requiring decisions regarding when it is appropriate and efficient, and when it is a detriment. 

While cognitive offloading to avoid learning tasks seems an obvious problem, little actual research exists to demonstrate the damage done. Some would argue that if technology can replace an activity and that technology is readily available, why bother to “learn” the skill in the first place? Why learn information if your cellphone can allow you to search for information when it is needed? Why learn basic calculation skills when you cellphone can also serve to do mathematical operations? There are responses to these challenges, sometimes offered by students or parents, but this analysis would take this post in a direction I did not intend. 

Here, I want to focus on learning to write and writing to learn by discussing a different learning task. This may sound unnecessary, but at present, there is a reason to take this approach. The justification for being indirect is that writing is a complex skill consisting of multiple subskills, and we learn to become competent at even a basic level over years and not weeks or hours. We are investigating an alternative to the traditional methods of instruction that can be subverted now, and we cannot rely on experience to help us evaluate and tease apart how the development of subskills are impacted. The insights and evidence of the potential damage done would take to long to emerge. As one perspective, consider the lingering impact of COVID on learning. What about the move to online learning did we not anticipate and what consequences are we still trying to mitigate?

AI in Learning to Code

Shen and Tamkin had an opportunity to investigate the impact of AI with adult programmers learning to make use of a new library. Think of a library as a collection of functions (tools to perform specific and commonly used tasks). Instead of having to write code to accomplish common tasks each time a programmer encounters a need, libraries allow programmers to call prewritten code snippets. It takes some work to make use of a library – what functions are available, how do you call the function you want, what inputs and outputs are involved and how are these integrated with the code you write yourself? The researchers recognized that the learning coders had to do to make use of a new library provided an opportunity to study how AI could help and hinder learning a complex process. 

Shen and Tamkin studied actual programmers as they worked to learn a new library. They suggested that the process be viewed as a tutorial including both background information and simple programming tasks. Programmers were assigned to a control and a treatment group, with the treatment group having access to AI. The learning phase concluded with an assessment evaluating multiple concepts and skills. Video of treatment group participants was collected to document how each individual used AI and worked on the programming exercises.  

The researchers found that the treatment groups did not differ significantly in the time spent learning, which they found surprising. On the post-test, the largest group differences were in debugging skills. Smaller skill differences were found for code reading and conceptual understanding. Those without access to AI made more coding errors on the practice tasks, spent more time practicing debugging, and ended up with better skills on the outcome evaluation. How AI was used differed greatly with some simply asking AI to solve the coding challenges and others who only asked higher-level questions of the AI tool. Some users had the AI tool solve the coding challenges and then retyped the solutions themselves (rather than copying and pasting). This was not an effective strategy. 

Generalizing from the coding study

I have spent considerable time both coding and writing and I have always found the processes to have similarities. While others may find this a strange observation, I have always said that coding and writing were the two professional tasks I learned I could not perform later in the evening if I wanted to get a good night’s sleep. Reading was fine. Grading was fine. Something about both coding and writing was cognitively stimulating, making it difficult to sleep. 

The application of AI to complex skills is interesting, but difficult to study. Clearly, a single skill would seem very unlikely to be developed if a learner could completely substitute AI for practicing the skill. However, it seems possible that learning a multiple-component skill such as reading or coding might benefit from replacing specific components with AI under certain circumstances. We have limited cognitive capacity and substitution for some components of a complex task could allow the remaining components to receive more attention until well learned.

Learning to write might represent an example. I have often referred to Flower and Hayes’ writing process model when describing the components of writing and writing to learn. The use of AI to offer content to provide the basis for a writing task and perhaps even to offer a structure to guide the organization of a writing product could free up capacity to focus on lower-level skills such as spelling, grammar, and coherent paragraphs. In contrast, I typically use Grammarly while I write to allow to move more quickly while relying on this AI tool to alert me to possible spelling and grammatical improvements. 

Part of what Shen and Tamkin observed in their qualitative observations of the different learner-imposed focus of AI and the relationship of differences to what was learned or not learned offers a related perspective. Debugging is an important lower level coding skill and having AI debug code appeared to limit a coder’s ability to debug when working without AI. 

Suggestions for Learning to Write and Writing to Learn

AI can support both “learning to write” (developing writing skill) and “writing to learn” (using writing to deepen understanding), but depending on which writing skills are the goal best practices should differ.

Learning to write: skill development

Here AI should be thought of as a coach, not a ghostwriter.

Emphasize feedback: Tools like Grammarly give immediate feedback on grammar, syntax, cohesion, and organization, helping students revise iteratively while concepts are still fresh.

Structure and separate subprocesses: Generative tools can help students brainstorm ideas, outline structures, or identify expectations for different types of writing (e.g., sample introductions, transitions).

Process?first policies: “Writing first, AI second” approaches ask students to draft independently, then use AI for critique and revision. When coders used AI in the Shen and Tamkin, this is the general theme that seemed most successful. 

Writing to learn: thinking with text

When the goal is conceptual understanding of content knowledge, AI is best used to amplify reflection, not replace it.

Clarifying concepts for the writer: Students can ask AI to reexplain readings, generate examples, or pose practice questions, then respond in their own words, using writing as a space to consolidate understanding.

Challenge personal understanding: AI can generate counterarguments, alternative explanations, or “what if” scenarios that students must address in writing, pushing them beyond summary toward analysis. Why do others disagree with the summary I am creating? What can I offer to support my position and what are the limitations of the alternative?

Shared design principles

There are some guidelines these goals for writing. Across both purposes, similar design choices matter.

Make process visible: Require artifacts – notes, outlines, draft histories, and brief process memos about when and how AI was used. Document the transition from any use of AI to products student has generated. 

Align AI roles with goals: For skills (learning to write), let AI focus on feedback, exemplars, and mechanics; for content learning (writing to learn), keep generative help outside the main composing space and treat it as a prompt.

Previous analysis of technology and the writing process

Sources:

Flower, L., & Hayes, J. R. (1981). A cognitive process theory of writing. College Composition & Communication, 32(4), 365-387

Shen, J. & Tamkin, A. (2026). How AI impacts skill formation. arXiv preprint arXiv:2601.20245 (this study has yet to officially be published)

One thing I miss as a retired academic is going to the office daily and having the chance to share ideas on common interests. My background was in educational psychology and topics such as AI and learning would not only be relevant to me, but also to the people I had lunch with and passed in the halls. I would have been interested in my colleagues’ take on the pros and cons of AI in classroom settings. Were they concerned about cheating? Had they encountered students who cheated and how did they know for sure that their suspicions were justified? Had they modified the assignments they had always used or perhaps abandoned these tactics as untrustworthy? 

I still find myself thinking about such topics and despite no longer having firsthand experience, I wonder what I would do should I still be working. When something is that important in your life and self-view, it doesn’t leave you, and I cannot help but continue to explore such topics and share my findings and opinions through outlets like this. 

Beyond the internet, AI poses a tremendous challenge, with both its opportunities and its risks. Cheating obviously falls in the risks category. It challenges how accomplishments are evaluated and the results shared with learners and other interested parties (e.g., employers, those involved in competitive selection processes for limited-enrollment programs, the instructor in subsequent courses). It also poses a challenge to our efforts to craft assignments we are confident will contribute to student learning. If tasks are not completed as we assume, we cannot trust the markers we use to evaluate what students know, nor can we rely on them to guide our decisions about when to move on and what we can assume will make sense in new instruction. 

Without my colleagues, I now must rely more on what I can read or find online to form my own opinions. This is a difficult and relatively recent problem and little I would regard as proven seems to be available. There is plenty of advice and personal perspectives and folks willing to offer books on the topic. I might as well offer my own perspective on a specific instructional situation, since, at present, ideas focused on specific tasks in a specific type of classroom are the only ones I don’t immediately find myself arguing with. 

The Opposite of Cheating

I have been reading The opposite of cheating: Teaching for integrity in the age of AI (Gallant & Rettinger). It is well written and well referenced, but the type of source I find myself both rejecting and applauding when it comes to specific recommendations. Typically, a negative reaction stems from the impracticality of a suggestion given my own circumstances. I doubt it is reasonable I should expect authors to create a master model differentiating when a specific idea can be applied as that would add too much complexity and readers need to be active participants in finding what they should take from a resource. Anyway, this book made a point that sparked what follows. 

Writing to Learn

Written products played a significant role in some of the courses I taught. I assumed the products were a) an incentive to read the sources I expected students to read and listen to presentations I and students made, b) a way to demonstrate understanding and depending on the assignment consider applications, c) a task that involved the student in thinking in ways that led to understanding and retention, and d) a way to evaluate students. Having students write in isolation is one of those common tasks that has come under suspicion because of AI. 

Back to the “Opposite of cheating”. One of the authors’ general suggestions is to evaluate the process, not the product. I wrote a post some time ago making a very similar point. I think it helpful to explain why emphasizing what I would describe as subprocesses allows not only what might be described as surveillance, but also a superior instructional approach. 

Similarity to the strategy of showing your work.

    Yes, a requirement in what is probably math classes that you show your work was partially a check on whether a student had done the work, but just as important it was a record of the processing that was involved. A student and the teacher had access to the student’s externalized thinking. This visible record might be used by the student when the process breaks down and must be adjusted. It also provided someone else the opportunity to follow the student’s logic. The concept of externalized thinking has many applications for those who propose that cognitive research is useful to educational issues.

My long term interests in showing your work have focused more on writing and a specific application of writing often called writing to learn. Given a writing to learn task, assigned by a teacher or taken on as a personal strategy, a student could, of course, simply start writing or feed a prompt to their AI tool of choice. Here again, the “show your work” strategy can serve as both a check that you have done the work and a benefit to deeper thinking. 

I have been influenced by the logic and justification of advocates of personal knowledge management and the second brain. These concepts, when considered carefully, are clearly process-oriented: engaging purposefully and thoughtfully in specific processes benefits the products they produce, and externalizing processes that could be performed internally enables them to be performed more skillfully. I have long been interested in the Writing Process Model (Flower & Hayes) and variations. These researchers sought to develop a model that identifies the processes of writing and how the processes interact to create a written product. One benefit they proposed for such a model was the identification of component skills, allowing more efficient development of proficiency in individual skills. Identification of processes could guide both the topics researchers pursue and the instructional practices relevant to the classroom.

Connection with the topic of mitigating cheating

Let me start with this claim: an externalization requirement can serve both the purpose of ensuring that a process has been executed by a person and the educational goal associated with the assigned task. I think this works well when the goal is writing to learn.

I already indicated that writing to learn (or learning to write) can be broken down into subprocesses. Rather than relying on the Writing Process Model, allow me to offer a simpler approach for this situation.

In order to complete a writing to learn assignment, a student must:

Read the content

Identified what she felt are important ideas in the content

Processed this collection of ideas to understand and apply

I assume this is acceptable as a gross level description. If an educator relies on only the product turned in, with AI the educator must guess whether any of these tasks had actually been performed by a given student.

Those of us who make use of Personal Knowledge Management tools engage in these processes even though we are not accountable to an educator responsible for our skill and knowledge development. We do these things because we believe they deepen our understanding and strengthen our ability to craft better products.

We integrate a variety of tools while we read that would allow someone else to agree that we have in fact read.

Most of these tools involve highlighting and annotation as part of the reading process. The highlights and notes serve as an external representation of what we regard as important ideas in the content.

We then extract highlights and annotations from the original context so we can store and manipulate these elements more effectively. Having these elements separated and independent allows their long-term access and allows further processing such as linking, tagging, and secondary elaborations to occur. We value this growing and ever-modifiable collection as what has become popular to describe as a second brain that can be searched and explored for new insights and the generation of products.

The tools we use are ever improving and the skills in using these tools are being constantly scrutinized in search of greater efficiency and effectiveness.

The tools are there and it is easy to find free options. There is long-term benefit in learning to use such tools as skills relevant to lifelong learning. Why not teach these techniques to students and use the potential side benefit of accountability?

Hypothes.is as a starting point – try it you might like it

I first used Hypothes.is because I was interested in social note-taking with students. Simply put, this perspective argues that there may be benefits to a system that allows students to share notes. What did others find interesting or valuable in an assigned reading, and what might comments they made in response to what they highlighted as important reveal that others may not have considered?

This same tool could be applied such that an individual’s highlights and notes be available just to the instructor rather than the entire class. This covers “was it read” and “were ideas I thought important identified? 

The process for exporting from Hypothes.is works like this:

How to Export Annotations

Activate Hypothesis: Go to the webpage or document you’ve annotated and open the Hypothesis sidebar.

Open Sharing Menu: Click the “Share” button.

Select Export: Choose the “Export” tab.

Select Annotations: Use the dropdown to choose which user’s annotations to export (your own, a specific group, etc.).

Choose Format: Select your desired file type (e.g. HTML, plain text).

Export: Click the “Export” button to download the file, or “Copy to clipboard”. 

A screenshot of Hypothes.is in use. Hypothes.is is a browser extension so the content must be something online or something you can open in a browser (e.g., pdf). The content window on the left is where the reader highlights, annotates, and reads. The highlights and notes appear in the column on the right. 

Organize and Elaborate

At this point, I would now bring individual elements into a tool such as Obsidian, which I would not hesitate to introduce to college students. This tool is designed to store a large collection of individual idea notes, tag them, create links among them, and extend individual notes by generating secondary notes (elaboration). I raise this tool as an opportunity, not because there are no other options. Perhaps this mention of this tool will raise the curiosity of those willing to go a little deeper. 

There are other basic ways to do this. In the next stage before writing, you might open a document in any word processing tool and copy and paste individual notes or ideas from the notes or highlights into this document. As you proceed, you might cut and paste from this working document to better organize topics and integrate them into your final product. Even with the many personal knowledge management tools I use, I often take this simple approach when approaching the final stage of a project. I might cut and paste chunks of text and citations from the content I have accumulated into a common document. Often, this is not just about collecting ideas from a single source but bringing together ideas from multiple sources. I open this “collection” document in a separate word processing window and work from this narrowing of material into a draft of the product I am creating. 

Some writing tools even offer visible ways to do this. For example, Scrivener provides notes as note cards that can be moved around in a space to explore organizational options. Even if you do not intend to use a tool like this, visualizing the approach may be helpful. The “corkboard” option in Scrivener is shown below. Here you can see how individual project-related notes have been moved to this corkboard. The notes can be dragged around to create an optimal structure.

Summary  

This post focuses more on a concept for discouraging AI cheating more than on a detailed tutorial for using the tools involved. The core idea is that tasks can be assigned that are both beneficial for applying subskills to the writing-to-learn process and useful for documenting students’ completion of these subskills. I have identified specific tools and tactics, but there are likely alternatives for everything I have used as an example.

Source

Gallant, T. & Rettinger, D. (2025). The Opposite of Cheating: Teaching for Integrity in the Age of AI (Vol. 4). University of Oklahoma Press.

The concept of generative activities has consistently shaped my thinking and teaching on learning. I admit that such activities are not ultimately necessary. Still, they represent ways for learners and those who try to help them grow to understand and imagine how skills and knowledge might be applied. Focusing on generative activities was particularly useful in my interest in studying – the work an individual does to make experiences personally informative and useful.

Basic Definitions:

Studying – the mental and external activities a learner engages in after exposure to potentially useful experiences that are intended to store a representation of these and create meaning.

Generative activities – external tasks intended to encourage productive cognitive (mental) behaviors 

Why is this perspective important at this time? My concern is that certain uses of AI are frequently being substituted for generative activities allowing individuals to accomplish tasks without achieving the cognitive benefits (i.e., retention, understanding) engagement with the generative tasks make more likely. 

Why would learners substitute AI for generative activities? It seems likely they see AI as producing an equal or even superior product without the effort required to create such products on their own.  This reflects both a focus on short-term benefits over long-term benefits and probably a lack of understanding of how personal knowledge and skills are developed, or perhaps even a disinterest in developing these personal attributes.

Some background:

When I explain the concept of generative activities I like to start with Rothkopf’s concept of a mathemagenic task because this researcher’s focus tends to make intuitive sense to most people. Rothkopf was interested in questions and variations in how questions might be associated with written material. 

Questions presented before you read. 

Questions presented after you read. 

Inserted questions – questions added within text. 

Different types of questions – application questions, factual questions. 

Mathemagenic tasks

The made-up word mathemagenic translates roughly as giving birth to knowledge, implying that in attempting to answer questions, you might accomplish something else – a better likelihood of future retention, greater likelihood that you would recognize possible applications – that would not have occurred without exposure to the questions. My favorite example relates to the challenge educators often face in encouraging students to see the relevance of general concepts they have been taught. This translates as connecting new ideas with what you already know. The examples and the relevance are potentially there if you can make connections. So, what not ask students directly – provide an example of XXX? If personal examples exist, but learners have not made the effort to make the connections, perhaps the request will encourage that specific cognitive effort. 

There is a huge body of research on all aspects of questioning. Questions are an everyday classroom activity, but the insight is just why do we spend the time, and could a more careful use of questions result in improved results? My favorite example here is what is called wait time – the average delay after asking a question (silence to allow thinking) is a little over a second. If we want students to think, typical behavior in classroom discussions is not particularly rationale. There is reason to examine and challenge typical behavior.

Anyway, questions are an external task that can be used to manipulate – change the odds of – productive cognitive behaviors. I suggest adding one important final point: a learner can ask herself questions, e.g., using flashcards. So various ways in which questions can be generated and used are an aspect of what those interested in study behavior investigate. 

Generative Tasks

The concept of generative activities is simply an expansion of this same idea and asking questions would be one of many generative strategies. The idea of generative activities is not new (Wittrock, 1974, 2010) and to many educators may seem obvious and a reflection of common classroom practices. While true, researchers have attempted to understand the underlying mechanisms and to consider just how efficient different activities were especially in the comparison of one to others ( Fiorella & Mayer, 2016). A personal interest and one clearly relevant to the topic of how AI is applied in classrooms is writing to learn. I have always felt through self awareness that requires careful examination of existing ideas and integration of ideas from a variety of experiences to produce a product. There is a substantial body of research to support such perceptions (e.g., Graham et al., 2020). To be clear, researchers consider a variety of writing activities under the umbrella of writing to learn. The product need not be a massive, semester summarizing paper, but perhaps also notes and short, five-minute end of class descriptions related to the content just presented. 

Caveat

One issue I think is important that may not be apparent in the notions that generative activities are intended to encourage productive cognitive skills is that such skills may occur without this external requirement and guidance and there is always the possibility that for some motivated and capable of thinking deeply, without such tasks, the task represents a form of “busy work”. In other words, the task adds little beyond annoyance. Of course, the reality is that educators in actual classrooms typically do not feel that they can arbitrarily assign tasks to some students and not others, so they must always deal with reactions to assignments, both legitimate and resulting from laziness. 

AI and Generative Tasks

AI discussions related to education always seem to generate a good news / bad news situation. There seem to be several examples that apply to this general topic.

AI can be applied to render the potential benefits of a generative strategy useless. For example, if AI is used to respond wholly to a writing-to-learn assignment, the learner completes the assignment without engaging in much cognitive work. The educator is then in a position of assigning a task that takes valuable learning time and adds a commitment to the effort to provide feedback, but has little impact. 

In contrast, AI can be used to formulate questions (both objective and open-ended) related to assigned material and to respond to a learner’s responses to such questions. Learners can even generate such activities on their own.  It seems to me that the use of what might be described as short essay questions offer a unique advantage that would be difficult or at least very time consuming for the educator to administer. AI tools are very flexible and can ask and react to the answers for different types of questions. Short answer questions are a form of writing to learn and involve greater “retrieval practice” benefits than formats such as multiple choice that are useful, but less demanding of retrieval. 

Summary

My effort here was intended as a way educations might frame their way of thinking about AI in classrooms using  examples I assume are familiar. I hope this approach can be generalized. Of course, the challenge is in manipulating AI-based and any assigned activities so that productive thinking activities are encouraged and also that students gain insight into the importance of the mental work that is required of certain task. I understand this may seem obvious, but the work of adjusting to the advantages and disadvantages of AI tools will take some time and careful study. For example, I wonder if writing and organizing notes may accomplish much the same benefits as creating a writing to learn product. Learning to write is somewhat different than writing to learn although writing across the curriculum offers a secondary benefit of practicing writing skills. There are plenty of options to consider. We presently do little to teach advanced note making skills and note using skills even though these topics have received a great deal of attention as benefits to out of school functioning. 

Citations

Fiorella, L., & Mayer, R. E. (2016). Eight ways to promote generative learning. Educational Psychology Review, 28(4), 717-741.

Graham, S., Kiuhara, S. A., & MacKay, M. (2020). The effects of writing on learning in science, social studies, and mathematics: A meta-analysis. Review of Educational Research, 90(2), 179-226.

Rothkopf, E. Z. (1970). The concept of mathemagenic activities. Review of educational research. 40(3), 325-336.

Wittrock, M. C. (1974). Learning as a generative process . Educational Psychologist, 11(2), 87–95. https://doi.org/10.1080/00461527409529129

Wittrock, M. C. (2010). Learning as a generative process. Educational Psychologist, 45(1), 40-45.

I found the Writing Process Model (Flower & Hayes, 1981; Hayes, 2012) helps me think about the development of writing skills and the specific application typically described as writing to learn. This model identifies the processes and the interaction among the processes involved in writing and has been used to guide both writing researchers and the development of instructional tactics. 

The model provides researchers and instructional designers with a concrete framework to work with by identifying specific skills that can be studied as the source of individual differences in writing skills or targeted for development, assuming that greater proficiency with these skills will lead to more effective writing.

I first used this model to speculate how specific technology tools could support writers. For example, technology offers powerful ways to take and review notes, a method for planning by creating an outline or “mind map”, and a way to record text that allows for easy manipulation and revision. I have begun to think about the model in a different way brought on by easy access to AI for writing in general and in classrooms more specifically. Given the general goals of learning to write and writing to learn, when do specific uses of AI facilitate and when do these applications harm the development of the writing subprocesses as involved in each category of writing experience? Put another way, instead of thinking of AI as an all-or-nothing approach to creating written content, would it make more sense to evaluate the use of AI in impacting writing subprocesses and perhaps have students use AI more selectively?

The Writing Process Model

The model identifies three general components a) planning, b) translation, and c) reviewing (see the following illustration). Planning involves setting a goal for the project, gathering information related to this goal, which we refer to as research, and organizing this information so that the product generated makes sense. The goal may be self-determined or the result of an assignment. Research may involve remembering what the author knows about a topic or acquiring new information. Research should also include identifying the characteristics of the audience. What do they already know? How should I explain things so that they will understand? Finally, the process of organization involves establishing a sequence of ideas in memory or externally to represent the intended flow of logic or ideas.

What many of us think of as writing is what Flower and Hayes describe as translation. Translation is the process of getting our ideas from the mind to the screen and this externalization process is typically expected to conform to conventions of expression such as spelling and grammar.

Finally, authors read what they have written and make adjustments. This review may occur at the end of a project or at the end of a sentence. Authors may also solicit advice from others rather than relying solely on their own review.

One additional aspect of the model that should not be overlooked is its iterative nature. This is illustrated in the figure, which presents the model using arrows. We may be tempted, even after an initial examination of this model, to view writing as a mostly linear process – we think a bit and jot down a few ideas, we use these ideas to craft a draft, and we edit this draft to address grammatical issues. However, the path to a quality finished product is often more circuitous. We do more than make adjustments in spelling and grammar. As we translate our initial ideas, we may discover that we are vague on a point we thought we understood and need to conduct further research. We may decide that a different organizational scheme makes more sense. This reality interpreted using our tool metaphor would suggest that within a given project we seldom can be certain we have finished the use of a given tool and the opportunity to move back and forth among tools is quite valuable.

This model describes the processes identified by Flower and Hayes, but ignores two other components. The first is the writing task, which consists of the assignment and any writing completed at a given moment. The other missing element is the long-term memory of the writer. The long term memory or existing knowledge provides a source of information and strategies that the writer can use without resorting to new research. 

Hayes (2012) updated this model to incorporate additional research and comments from colleagues and his own laboratory. Later work placed a strong emphasis on both revising the planning component and subsequent drafts, as well as on the role of working memory. Working memory recognizes that cognitive capacity is limited, meaning activities must fit within the existing capacity or perhaps be ignored. Processes become less demanding less capacity as a function of practice and aptitude and among other topics, is vital in understanding changes that could occur for learners across the grade levels. For example, Hayes noted that keyboarding is more difficult for younger learners than handwriting and products produced on a computer would thus appear of lower quality. Differences in cognitive demands do change with experience with keyboarding, eventually becoming less demanding than handwriting. Other factors can also impact cognitive demands and in one example, Hayes noted that studies requiring adults to write in all caps reduced the quality of the final product because transcription in this form is less practiced. As an aside, I could not help recognizing the tendency of some to write in all caps on social media and my perception of the quality of what is produced in that format. Anyway. Finally, Hayes recognized that writing was a motivated activity and differences in motivation could emphasize the processes.  

One of the significant patterns in writing proficiency Hayes and others (Bereiter & Scardamalia, 1987) associated with the limited capacity of working memory and the gradual development of proficiency in the cognitive writing subskills, was the tendency to move from what Hayes and others describe as knowledge telling to knowledge transforming. The first approach results in a dump of ideas from long term memory triggered by the assignment and information that has already been recalled. In other words, a product can be generated without changing what is stored in memory or how this content is organized. Most instructors have a feeling for how this works. They have asked students to reply to an essay question and received at least some responses that seem to be everything students knew about the topic rather than a specific answer to the question. It can be hard to know if the student thinks they have answered the question or if this is just a frequently productive ploy. It is the first situation that shares characteristics with the idea of knowledge telling.

In contrast, writing classified as knowledge transforming requires that previously stored information be reorganized, reinterpreted, or extended based on speculation or insights. The writing task sometimes determines the difference, but transforming is more demanding and, when expected, is more likely to be produced by more capable writers. Writing to learn could involve either output, but the greater manipulation of ideas in knowledge transformation reflects the most significant benefit. 

Klein (1999) offers a review organized around the writing processes hypothesized to be responsible for learning. A brief summary of some insights from this paper provides examples of tasks and skills that differentiate knowledge telling from knowledge transforming, also relating the distinction to the processes involved. 

1. Point of utterance (no revision and limited planning). This explanation assumes that learning occurs in the attempt to generate comments on the topic (no revision or planning is expected). This category might be described as spontaneous writing – the learner is asked or personally commits to writing on a topic. An activity fitting within this category would be the five minute writing tasks some college lecturers assign at the end of a class. 
2. Genre-related – This explanation focuses on the benefits of transforming ideas to the structure demanded of a specific genre. For example, in a “position paper,” a writer is expected to take on of several possible positions are generate an argument supporting this selection. In selecting, organizing, and connecting knowledge to fit the demands of a formal writing task, the learner creates an understanding that would not exist without the imposed task.
3. Backward search – This explanation assumes that skilled writers formulate complex goals (characteristics of the desired product, audience needs, etc.) and then rework existing knowledge in terms of these goals. This seems a more generalized version of the “genre-related” explanation, requiring more sophisticated and complex problem-solving.

My own writing is generally linear with backtracking depending my assessment of how well things are going. This impression is based on my own behavior and may not accurately reflect the ideal recursive approach. When I take on a project, I assess what I already know and have notes on a topic and then read and take notes on additional material. I generate a rough idea of how these ideas could be organized (some would create a formal outline) and begin developing a draft. Often, I realize I need to fill a hole in my mental outline or find a reference in support of what I am trying to accomplish and have to take a closer look at my large collection of digital notes or read and take notes on another source or two. Upon completion of this initial draft, I reread what I have down to take care of lower level deficiencies (spelling, grammar) and often to add a little more material here or there so the document makes more sense or hangs together better. When writing for myself and without an external reviewer, I tend not to make major structural revisions. However, when I submit a paper for review, I do sometimes have to make larger adjustments even if the content is deemed useful. 

Where does AI play a role? Where should AI play a role? I have worked through my thoughts on these questions based on my own circumstances and also on what should be the circumstances for those in more formal learning settings. I want to produce content that meets a reasonable standard of quality, but I am not that interested in becoming a more accomplished writer. I want what I write to be a reflection of my experiences and what I have learned, and I assume I can learn from the process of writing. I am willing to invest time in this objective. I propose that others generate a similar analysis for themselves based on personal goals or goals for others they may be responsible for educating. 

How do I use AI? I primarily use AI to facilitate the research and planning components of my writing process and to perform some of the revision tasks. These emphases are consistent with my desire to learn from writing and my lesser interest in improving my writing skills. I am not suggesting everyone apply my priorities, I am suggesting it is possible to identify priorities and use AI strategically and efficiently.

Here is how this works. I have developed a large collection of notes consistent with methods of personal knowledge management and a technology-enabled second brain. This is a multidecade-long process that involves reading widely in the educational technology and educational psychology literature. When I get an idea for a writing project, I use AI to query this body of content for ideas related to my intended project. Based on the content I produce and queries for related information (some generated by AI), I review this new content and supplement my notes on the topic. I then often use AI (typically NotebookLM) to provide a structure for the intended project based on my collection of notes. I consider this proposed approach when generating my initial draft.

I use Grammarly heavily when I write. I use the pro version so I get constant feedback not only on spelling and basic grammar, but also more substantive recommended changes at the paragraph level. Grammarly identifies issues I should consider and offers suggestions. I suppose this could be a valuable learning opportunity, but I admit I just select the recommend changes that sound good. Grammarly has just pushed a major update that offers even more capabilities, but I have yet to explore which might be useful to me. 

The use of AI for research, planning, and text revisions do not limit my writing to learn. I think it is quite reasonable to associate AI capabilities with specific components of the writing process model, as applied to writing and writing to learn, in order to accomplish tasks that are not essential to developing writing skills or learning opportunities. The stickier problem presents itself when you must consider how to control the use of AI. 

Sources

Bereiter.C. & Scardamalia, M. (1987). Two models of composing processes (pp. 1-30). In C. Bereiter & M. Scardamalia (Eds) The psychology of written Composition. Erlbaum. 

Flower, L., & Hayes, J. R. (1981). A cognitive process theory of writing. College composition and communication, 32(4), 365-387.

Hayes, J. R. (2012). Modeling and Remodeling Writing. Written Communication, 29(3), 369-388. https://doi.org/10.1177/0741088312451260

Klein, P. D. (1999). Reopening inquiry into cognitive processes in writing-to-learn. Educational Psychology Review, 11, 203-270.

I was excited when Google announced its investment in developing NotebookLM as a tool for learners. First, I encountered the description of its new program Learning Your Way, which promised to personalize textbooks in various ways, including explanations based on personal interests. Examples of what this would look like were provided, but to submit your own content for the “learn your way” treatment, you first had to indicate your interest and then wait to be selected. I am still waiting. However, Google has also announced enhancements for the existing NotebookLM that make several proven study strategies more straightforward to implement, as well as a couple of unique ways to repurpose instructional text as a podcast or something resembling a slide-based lecture. The updates available to all are what I will describe here.

Before anyone gets overly excited, what Google can presently make available requires that an instructor or a learner must be able to load content into NotebookLM. There is no legal way in which what I will describe can be done with your existing commercial paper or even digital textbook. Companies producing instructional materials may either collaborate with tech companies or develop similar capabilities, but they are not doing so at present. I wrote a couple of textbooks earlier in my career. When my collaboration with a textbook company ended, I was given complete control of the copyrights and now I have full-length chapters I can use for demonstrations. This is what I am doing here. NotebookLM notebooks can be shared and I will make the content on which this tutorial is based available for you to explore. 

For those unfamiliar with the newest updates, NotebookLM presents as three vertical panels. The leftmost panel shows the content you have uploaded to NotebookLM as well as provides the opportunity to upload additional content and open individual files for viewing. The middle panel is where AI prompts are applied and the output from such prompts are displayed. The rightmost panel is where the older and new AI tools are activated and output from these tools can be accessed for later display. 

As soon as NotebookLM is launched, the middle panel displays a summary of the source files that have been added. Loading source files does not include them in this summary unless they have been selected (see check box). NotebookLM was developed as an RAG (retrieval-augmented generation), which means the prompts are applied to the content you provide rather than applied to a massive corpus available to the model. The new “Discover” tool available from the top of the sources panel breaks with this approach and allows a user to request additional information intended to be related to the purpose for a particular notebook that can be reviewed and added as an additional source. 

Before I continue, I want to identify one frustration I have with the way NotebookLM fits into my workflow. Developers have specific applications in mind when creating applications, and their vision may not include what an individual user prefers. My preference is to read, highlight, and annotate content before using an AI tool. While NotebookLM allows access to sources, it does not support highlighting or annotation, nor does it display such added layers from uploaded material. The workaround is to export personal additions and add this content to NotebookLM separately. 

Mindmap

New capabilities appear in the right-most column. The audio overview (podcast) and Mindmap capabilities were already available and will not be covered here. 

Quiz

The new tool that first caught my attention was the quiz. I have explored various ways to have LLMs ask questions about designated content, but having a built-in tool for doing so is unique. The tool generates a double-digit set of multiple-choice questions that are then stored as an associated resource. I have no idea why the number of items varies. When selected, the MC questions are displayed one at a time (see the following image) and then, no matter which alternative is selected, a comment is provided. The “explain” button (see red box in the following image) shows the content from the text associated with that question. If you want more questions, the set generated can be deleted and a new set generated. 

I think it is valuable when using this tool to be aware of your reaction and recognize you are working with an AI tool. As someone who had to generate multiple-choice questions for actual student exams, knowing what I know about student reactions, I would not use some of the AI-generated questions on an examination. I thought the stems of some AI questions were kind of vague or misleading, but the explanation clarified the logic justifying a given answer. If you understand this and are using the questions as a study experience, the ambiguity is probably fine. As a final comment to this post, I will explain why I prefer to ask an AI tool to generate open-ended (essay) questions on a specific topic and then wait to evaluate my response. Interacting with content in this way is more helpful. 

Flash Card Tool

Flash cards are a useful tool for implementing memory enhancement strategies such as retrieval practice and interleaving. While relying completely on flashcards would not be wise, this approach is effective and useful for quick reviews spaced over time. Again, the “Explain” button (see the second of the following images) offers a connection to the original content and this provides potential opportunities to review beyond simply verifying that a response is correct. Recognize that NoteBookLM generated 69 cards for the targeted chapter, so this tool covers the content in detail. 

Video Overview

The video overview offers what seems like an extension of the existing audio option. It strikes me as the equivalent of a lecture based on a series of slides. So imagine a large lecture environment in which you see the screen on which an instructor is projecting her PowerPoint and you are hearing the audio but not seeing the presenter. Having written this chapter, it is an interesting experience watching a lecture based on your words. It is possible to download the video so I decided to offer this download through YouTube if want to see an example. 

Just Use Chat

NotebookLM, before these built-in tools, was a way to interact with uploaded content using AI chat prompts. I find this capability more powerful than the tools that seem to recreate standard study experiences. The familiar tools are probably a way to engage less experienced learners, but I would argue that concepts such as debating and tutoring offer opportunities that take advantage of the language capabilities of LLMs. The following is an example of a chat I had with the content.

[Prompt] Ask me a short-answer question about generative tasks and wait to evaluate my reply

According to the sources, what two specific categories of generative activities are highlighted for their value in improving understanding, retention, or transfer, and what general type of content processing is involved in these activities?

——————————————————————————–

I will now wait to evaluate your reply.

[Learner reply] I remember the chapter talked about responding to questions and I can’t remember another example from the book. How about writing to learn, but I don’t think that was in the chapter.

[AI response] That was a strong attempt, and your intuition about “writing to learn” was very close to one of the specified terms!

Here is the answer to the question, based on the specific terminology used in the sources:

The two specific categories of research emphasizing generative activities that improve understanding, retention, or transfer are labeled as:

1. “Authoring to learn”

2. “Teaching to learn”

These activities involve a “generative processing of content”.

Your suggestion of “writing to learn” is directly supported by the text, as “authoring to learn” is described as involving the preparation and implementation of writing tasks. This generative processing of content is suggested to be superior in many ways to studying only for personal use.

While responding to questions is certainly a critical activity discussed in the sources—often serving as an external strategy for evaluation and regulation that compensates for weak metacognition—it was not one of the two specific research categories given the label “generative activities” in this context.

[ Added comment] What is not visible in my copies of this text from this interaction are the inserted numbered buttons that provide access to sections of text from the original text.

---

Notebooks can be shared. You are invited to explore the Notebook that was the basis for this prompt.

In K-12 education, writing often takes a backseat to other academic priorities. Elementary education is dominated by math and reading, while secondary education spreads its focus across a wide range of subjects. Unfortunately, this distribution of attention has left the development of writing proficiency hindered by several factors.

Limited Time for Writing Practice

One significant issue is the limited time students spend on writing activities, both in and out of the classroom. Research shows that only about 25% of middle school students and 30% of high school students meet the recommended minimum of 30 minutes of daily writing practice. Writing assignments at these levels are often brief, typically a paragraph or a short essay, with few opportunities for more complex projects that require synthesizing ideas from multiple sources.

While initiatives like “writing across the curriculum” aim to increase writing opportunities and integrate writing into other subjects, many teachers outside of language arts lack the training to effectively incorporate writing into their instruction (Picou, 2020). This lack of preparation, combined with time constraints, contributes to disparities in writing proficiency, as reflected in differences in NAEP scores across schools (Mo & Troia, 2017).

The Broader Benefits of Writing

Writing is not just a skill—it is a multifaceted process that integrates numerous subskills and offers significant cognitive and academic benefits. However, these benefits are only fully realized when students receive meaningful feedback on their work. Unlike other disciplines, evaluating written work is particularly time-intensive for educators, which may discourage frequent and substantive writing assignments.

Beyond skill development, writing also serves as a powerful tool for learning. Often referred to as “writing to learn,” this process involves cognitive demands that enhance understanding and retention (previous post). Writing tasks act as generative activities, externalizing thought processes and encouraging deeper engagement with the material. For example, when students are asked to provide personal examples of a concept, they connect prior knowledge to new ideas, fostering meaningful learning.

Two types of writing tasks stand out in their educational value: writing to explain and writing to persuade.

1. Writing to Explain
   This task requires students to learn something and then articulate their understanding through writing. The act of externalizing knowledge serves as a form of self-assessment, revealing gaps in understanding and prompting further learning. This process, often linked to metacognition, helps students refine their knowledge as they work to organize and express their ideas. As educational psychologist Graham and colleagues (2020) note, writing to learn has consistently been shown to enhance academic outcomes.
2. Writing to Persuade
   Persuasive writing involves crafting a position, supporting it with evidence, and addressing counterarguments. Despite its importance in developing critical thinking and reasoning skills, persuasive writing accounts for only about 20% of writing-to-learn tasks. This is a missed opportunity, as persuasive writing offers concentrated practice in analysis and argumentation, much like debate, but in a more efficient format for classroom use.

The Impact of AI on Writing Development

While writing is already underutilized in education, the rise of AI tools presents a new challenge. Educators are grappling with how to integrate AI productively without further reducing the time students spend writing. If AI tools are used to complete writing tasks for students, the generative benefits of writing—such as critical thinking and cognitive engagement—may be lost.

As someone who uses AI tools daily, I recognize their potential to enhance productivity and creativity. However, I also understand the risks. For example, I use AI to interact with a personal corpus of notes, allowing me to explore ideas before writing. This approach complements my writing process rather than replacing it. Students, however, may lack the motivation or understanding to use AI in similarly constructive ways. Under time pressure, they may rely on AI to bypass the cognitive effort required for writing, undermining the development of essential skills.

Addressing the Challenges

One strategy that I think would address both writing challenges would require an increase in supervised classroom writing. Such tasks could be improved with collaborative writing activities that included peer editing and revision. The peer responsibilities would include attention to both writing quality and content accuracy when the task is a writing across the curriculum task.

Moving Forward

The challenges facing educators are undeniably complex, and the rise of AI adds another layer of difficulty. However, ignoring these realities will not improve the situation. Writing remains a critical skill, both as a standalone competency and as a tool for learning across disciplines. By increasing classroom writing opportunities and leveraging collaborative approaches, educators can help students develop the skills they need to succeed in an AI-driven world.

I welcome your thoughts on this analysis and any ideas you might have for addressing the interconnected issues of writing development, AI integration, and educational priorities.

Sources

• Graham, S., Kiuhara, S. A., & MacKay, M. (2020). The effects of writing on learning in science, social studies, and mathematics: A meta-analysis. Review of Educational Research, 90(2), 179-226.
• Mo, Y., & Troia, G. A. (2017). Predicting students’ writing performance on the NAEP from student- and state-level variables. Reading and Writing, 30(4), 739–770.
• Picou, A. (2020). Are schools making writing a priority? New study shows students are not spending enough time. The Learning Agency Lab