I have resisted the complaints that reading from a screen leads to poorer comprehension for some time. Much of my own research and writing have promoted the educational benefits of technology and the multiple studies and meta-analyses reporting that comprehension is adversely impacted when reading on a computer or tablet seemed a challenge to my message. A recent post on Reading Rockets by Timothy Shanahan identified several new studies I had not read and encouraged another look and another post.
When I first wrote about educational applications of technology, my approach was strongly influenced by David Jonassen’s mid-1990s book Mindtools. Rather than focusing on programming, computer literacy, or drill software, Jonassen proposed that educators consider how using productivity tools (e.g., word processing) could change how learners explored ideas. Again using word processing as an example, how writing to learn with a computer might expand the cognitive benefits of the writing experience beyond that allowed by pencil and paper.
A more recent version of the “tool perspective” (e.g., Etchells, 2024) offers a more nuanced view and notes that tools are in many ways neutral and users determine whether tools will be used effectively or detrimentally. The ills associated with technology are challenging because technology offers opportunities and challenges. However, Etchells proposes that the challenges are not due to addictions or biological triggers. The challenges are often best described as bad habits. This creates a difficult situation for some with the tendency to promote but also allows others to blame the tool as bad and ignore the control and responsibility of the user. With a specific focus on reading from digital devices which of these perspectives makes the most sense? If I maintain a position of advocacy, can I focus on the responsibility of the learner in a way that offers potential solutions rather than blaming the victim?
My “out”
First, I have no concerns about my own nearly total reliance on reading from digital devices. Even if there is a minor hit to my comprehension, my digital reading is part of a process of getting from a review of the ideas of others to my own written output. Digital reading allows me to efficiently highlight and annotate and export these personalizations of my understanding into external storage that I have accumulated over the years and use months and sometimes years later as background for what I write. This is one of those situations in which I cannot direct others to research evaluating the relative effectiveness of this approach to knowledge work, but trying to approximate my process on paper seems pointless. Put another way, reading comprehension is just one component in combination with long-term storage, multi-document processing, organization, retrieval, the flexible connection of stored elements, etc. that are involved in the writing and teaching I do.
Shallow Comprehension
I am starting to find the redundancy of the basic research studies and meta-analyses that have filled journal space in recent years to be mildly irritating. Replication is important in science, but at some point, it is time to move on to replication and extension. The studies that are attempting to study the process of comprehension when reading from a screen or paper are an example of what I mean by extension. Because it is not obvious why reading text on paper or on a screen involves the cognitive processes involved in reading, repetition of differences in the product (answers to test questions) has a decreasing value.
It would seem that the existing research demonstrates an issue, but not an explanation. Reading from a screen and paper would seem to involve the same cognitive subtasks, but the results are not equivalent. The explanation that seems to have gained the widest acceptance (my judgment) is commonly called the “shallowing hypothesis. I attribute this claim originally to Nicholas Carr who wrote a popular book titled “The Shallows”. As I understand the argument, the reading we tend to do online mostly involves what we might commonly describe as skimming. This experience leads to a bad habit that is activated when we read material from a screen even with content we might read more carefully under other circumstances. The reading habits of doomscrolling are proposed to generalize to reading books.
Eye movement research offers a way to watch the process of reading. We know that reading on a screen moves faster and readers overestimate their understanding level. Speed to complete the reading of a passage and requests to estimate the level of understanding by asking for prediction of performance areas variables that are easy enough to collect. I studied the phenomenon of failed reader understanding of their level of understanding and recognized the issue as a problem of comprehension monitoring resulting in poorer calibration. Calibration is simply the accuracy with which you can estimate the quality of a future performance. I always explained it as the decision that a student makes the night before an examination. Can they stop studying now and go out with friends or should they keep working to get the exam score they want? I was interested in calibration as a challenge for less capable readers reasoning that simple fix-up problems such as stopping and rereading sections that were poorly understood was the basic adjustment less capable readers were unable to make. In reflecting on the challenge in the context of college student study behavior, I tended to think of the problem as a matter of efficiency. Students used to tell me they had read chapters several times in preparation for an exam. Rereading is a common strategy among the motivated, but less capable learner. It would be far better to recognize the topics you did not understand and focus on these topics. It would be even better to notice this immediately while reading and while being more aware of the context of the material being read. This would be the ideal time to reread a paragraph or two.
When I was actively involved in this research, this challenge was sometimes studied by inserting different types of errors in text and determining if readers could identify the problems. Create circumstances that should result in a failure of comprehension and see if a reader would recognize the problem. For example, a statement that existing knowledge should indicate was inaccurate or a conflict between two sentences in the same sentence. Better readers were better at identifying these inserted errors. The technique was criticized because it placed readers in the role of proofreaders which is not the way we actually read. At the time, I tried to investigate the issue of monitoring skills by using a similar technique without informing the participants in my research and using eye-movement recordings to determine what happened when readers encountered what should have been confusing sentences. At the time the technique was tricky to use and intrusive as readers had to maintain their heads in a fixed position by bracing their chin on a bar. My research traded one form of artificiality for another.
The sophistication of the equipment has drastically improved and eye movement techniques are being used to compare participants reading from printed material and from a screen. For example, a recent study compared 8th graders in Norway (Jensen and colleagues, 2024) reading items from the national reading assessment exam from paper and screen while recording eye movements. As part of the preparation for the actual experiment, the researchers asked about their experience with reading from paper and screen and determined that many students proposed they comprehended better when reading from screens.
The reading examination itself involved passages of text each followed by several questions. Readers could refer to the written material when answering the questions. The eye movements indicated that readers referred back and forth between text and questions more times while reading from the screen to answer the questions. The researcher found that the readers had greater difficulty understanding when reading from a screen. This makes sense, but the readers still performed more poorly on the questions which would seem to indicate they were aware that they did not understand and could not completely remedy the difficulty.
Solutions
I don’t see online reading as a behavior that will disappear or diminish in frequency. I wonder about the hypothesis that poor habits are developed from heavy exposure to the scanning behaviors that many apply in texting and scrolling through social media feeds such as Twitter, Mastodon, and Blue Sky. I admit my perspective differs because I am a reader of books and journal articles on digital devices. I try to remember that my age means this style of reading (long form) predated any exposure to reading on a device. So my habits could be different allowing strategies of reading to differ between long-form and the types of text more suited a superficial approach.
Shanahan proposes that if technology environments encourage maladaptive behaviors why couldn’t they be engineered to encourage more effective behaviors? I think there are already such opportunities with longer-form content, but the opportunities are both ignored by most readers and not taught in educational settings. The exporting of annotations and highlights I rely on is seldom a part of the reading process of most learners. Here is another example of the transfer of a habit. Younger students are not allowed to highlight or write in their print books and the highlighting behavior of college students has frequently been criticized. I have written extensively about a variety of ways in which educators and learners can increase their active engagement with digital content. I call this collection of techniques layering because the strategies are external to the text itself and yet provide opportunities for processing text (and video) in more cognitively active ways.
Maybe improved awareness is also important. If skimming can be understood as a habit sometimes applied inappropriately in some settings, greater metacognitive awareness of goals and content would seem to be important. Concepts such as deep reading versus recreational reading are not new and the concern for shallow reading on devices is just another version of this older way of thinking about what learning from reading requires.
Sources
Carr, N. (2020). The shallows: What the Internet is doing to our brains. WW Norton & Company.
Etchells, P. (2024). Unlocked: the real science of screen time (and how to spend it better). Hachette UK
Jensen, R. E., Roe, A., & Blikstad-Balas, M. (2024). The smell of paper or the shine of a screen? Students’ reading comprehension, text processing, and attitudes when reading on paper and screen. Computers & Education, 219, 105107.
Jonassen, D. H., & Carr, C. S. (2020). Mindtools: Affording multiple knowledge representations for learning. In Computers as cognitive tools (pp. 165–196). Routledge.
Singer, L. M., & Alexander, P. A. (2017). Reading on paper and digitally: What the past decades of empirical research reveal. Review of Educational Research, 87(6), 1007–1041. https://doi.org/10.3102/0034654317722961.
I came to my interest in PKM as a retired educational psychologist with research interests in how students might take and use notes more effectively. This connection led me to translate the popular books on note-taking, second brains, and secondary topics such as whether notes should be taken on cards using a pen or on a computer into the descriptive models of note-taking and studying and research that were already familiar to me.
What I missed with PKM was the applied research I knew was so common in answering questions and suggesting ways to improve classroom-based note-taking. I had my personal experiences to provide some insights. Academic researchers use what I learned could be called PKM techniques in their profession. They read books and journal articles and take notes at academic conferences. They highlight and annotate and for years have used tools such as Endnote to track the references they have accumulated over years of work. Most of us just did these things without guiding models or books focused on “the process”. Even those of us who studied student note-taking were likely oblivious to the possible connections to our own daily behavior. We didn’t discuss related topics with colleagues or share recommendations for new reference managers. My personal experience in recent years is still an exception. When I talk with colleagues about my new interests and describe permanent notes or building a second brain their eyes glaze over. These folks are making a living with merit money depending on publication frequency and grant money secured and they are mostly oblivious to proposals for better knowledge management. I keep their reality in mind and it does make me wonder.
I know colleagues who have authored hundreds of publications. The Zettelkasten fans who point to the prolific publication record of Nikolas Luhmann have nothing on these scholars. Plus, the individual I have in mind collects the data necessary for these publications in addition to just writing. They do work with students so their situation while common may be a little different. However, to justify an approach anecdotally with Luhmann as the most common example opens a position taken to counter examples. There are many prolific counterexamples when it comes to knowledge workers.
The Knowledge workers I know have a body of information they have mastered and they tend to use this body in combination with just-in-time learning of material related to a topic that has caught their attention. They read constantly in narrow areas related to their work and the focus of this reading drifts about depending on the research that is their present focus. The component of PKM they tend to ignore is a way to connect over an extended period of time with novel associations that might lead to creative insights.
I do find some of the ideas and methods proposed by PKM and second brain advocates to be interesting. The point I am making concerns the lack of actual evidence regarding the proposed advantages of these techniques.
A category system for understanding and applying learning
I have come up with a potentially hierarchical system for the sources of information concerning learning. The items on this list differ in a variety of ways and different individuals likely value these levels to different degrees. Learning must be based in biology, but the success of a recommendation does not necessarily require a neurological basis. You can simply test a method against alternatives and determine which is most effective. Sometimes such tests at one level of my proposed hierarchy can be used to justify strategies at a different level. I think this is the case in using basic cognitive and memory research to provide insights into strategies to be tested at the applied cognitive strategy level and the applied cognitive strategy level has been used to justify strategies at the PKM and Second Brain level. However, speculative justification is not the same as actual experimental justification at the higher level. This is the issue I have with PKM and Second Brain proposals. As you move up this hierarchy, the complexity of the circumstances of application increases, and elements of this complexity may impact outcomes in ways that are not an issue at lower levels where these variables can be controlled. Things that work in the carefully controlled environment of the lab do not necessarily work outside of the lab. This does not mean the lab results are not real. It just means proposed applications need to be tested to see if potential and often unknown interacting factors can be ignored
Levels:
Personal Knowledge Management and Second Brain Strategies
Cognitive Note-taking strategy research
Basic Cognitive Memory and Comprehension research
Neuroscience
PKM strategies
For those who have read one or more of the personal knowledge management “how to do it” books, I will identify some of the recommended strategies I think should be tested. For others, I will try to include a few additional comments to offer a short idea of what these strategies recommend.
The notes intended to be retained should isolate specific concepts or ideas with enough context that at a later time the note taker or someone else with relevant background could understand the meaning of the note without access to the original context that prompted that note. Isolation of specific, meaningful ideas in storage is an important attribute of this strategy. (e.g., Atomic notes, permanent notes)
Immediately and over time, these isolated notes are to be connected with other notes (links). The connections are to suggest ways in which ideas relate to each other as similar, contribute to a process, or are part of an argument. Systems argue that important notes should immediately be stored as part of such relationships and new relationships can be added as they become apparent to the note-taker.
Individual notes are often also grouped by the addition of one or more additional tags common to other notes for which the tags are appropriate. Tags are an important way to search for and retrieve related notes.
Revisiting and reviewing are expected. Exploring the collection of notes as part of the process of developing the value of the stored information is an end in itself, not just something done when using stored information to accomplish a specific project. The review, which is sometimes random, increases familiarity and allows for the addition of new tags, links, and notes.
Note-taking and note-use are long-term processes in which notes are not necessarily taken with a specific purpose in mind. Tags, links, and other organizational strategies can be used to find relevant information when a novel purpose emerges.
My goal is not to challenge whether these tactics are beneficial. I am of the opinion that any external activity that results in analysis, decision-making, personalizing through summarization, or self-evaluation deepens understanding and improves retention. Simply put?—?thinking is good. A better way to explain my interest is whether the specific strategies I have mentioned are an improvement in comparison to what I would describe as a control condition. For example, do these tactics offer an advantage over the way say students traditionally take free-form notes and then review these notes from time to time? If the answer is yes, we should be emphasizing and teaching the PKM strategies to high school and college students. I think other control conditions may also be important. For example, in a recent post, I suggested that AI allows a learner to “chat” with stored notes and highlights and that using this means of interacting with content may have advantages over “manual” strategies emphasized by PKM enthusiasts. This approach might render many of the PKM strategies busy work.
As is so often the case with research, the investigation of one issue leads to the need to investigate related assumptions. For example, many have now concluded that reading from paper is advantageous to reading from a screen, and taking notes by hand is advantageous in comparison to taking notes with a digital device. However, with long-term goals for learning in mind and depending on the comparisons I have proposed, wouldn’t having stored information in a searchable and AI-accessible form be essential?
I wonder. There are so many areas in which developments and thought experiments can move faster than the research that would support or reject advocated practices. Ideal tests of some of the issues I have identified probably will not happen. Part of the rationale for PKM is to create a system for improved use of the learning experiences we have over long periods of time. Longitudinal research is expensive and fraught with problems of maintaining a commitment to treatment alternatives. How important are these issues and are they important enough to invest time and money? However, ideal approaches aside, I cannot find research that even considers the proposed benefits of the PKM component strategies I have identified over a short period. I think some of the issues I have identified warrant at least that level of commitment.
How much of the “building” part of building a second brain is necessary? With the opportunity to direct AI at accumulated resources, how much of the linking, tagging, reviewing, and re-expressing notes is productive? Some version of this question has long bothered me. I studied classroom note-taking as an academic and became familiar with the massive research base associated with different ways to take and use notes, the cognitive and metacognitive processes associated with activities in different note-taking approaches, and individual differences in the capacity and motivation to execute specific processes effectively. While there is a great deal of speculation, multiple books, and the sharing of numerous personal processes in online posts, where is the research on PKM and Second Brains? The research on learners taking and using notes in academic settings may offer a general structure, however there would seem important differences in the use of self-directed note-taking outside of an environment that involves processes that must accomplish goals set by others to produce products or perform tasks after short time intervals (the time between exams or writing assignments).
The terminology is different. You do not find classroom-oriented research considering fleeting versus permanent notes or the ideal types of tags and links to organize notes and prepare the user for future goals and tasks that often do not exist when the notes are taken. I try to translate the proposed strategies outlined in popular books for the audience that takes notes to benefit them under these circumstances (Ahrens, Doto, and Forte). Can the research on generative cognitive processes, metacognitive comprehension monitoring, and retrieval practice serve to determine if the recommendations made by these writers are reasonable? Perhaps the strategies for creating different types of notes (fleeting vs. permanent), strategies for linking and tagging, and reviewing to find new associations are just busy work.
Does AI provide an alternative?
MEM.AI was the first note-taking and storage tool I used that came with an embedded AI tool. The promotional material I read before investing time and money into MEM proposed that the availability of the embedded AI offered an alternative to the tags and links in a tool such as Obsidian. MEM.AI allows the manual assignment of something similar to tags and bi-directional links. The embedded AI did a couple of interesting things. As you built up a collection of notes, the AI offered suggestions based on the material you have accumulated. Similar notes were identified based on common elements of a new note and existing notes (see red box in the following image). You could link to suggested notes if you found the suggestions of value. Tags for a new note were proposed beyond any rags that were manually added (see red box in the following image) and any of these suggestions can be accepted. I have recently written about multiple ways note-taking systems can be augmented with AI, but MEM.AI is built with some of these processes assumed rather than potentially available through the creation of clever prompts. MEM.AI proposes that it creates a self-organizing workspace which seems a reasonable description.
MEM.AI has one more unique feature that seems a powerful use of the self-organizing workspace concept. There is a built-in word-processing capability called Smart Write. When you enter a forward slash / within any existing note or within an empty note, you can then enter a chat prompt. You can ask a question or make a request. With an existing note, the AI extends what already exists in what are often interesting and creative ways relying on what appears in the note you are working on and the corpus of notes you have saved. The output appears within that note and can be copied and pasted elsewhere or left as part of that note. It can be argued that this is a way to discover connections to other material more directly than what might have been manually indicated by creating a link to another specific note.
Here is an example.
The following (Permanentnotes) is a note I created with two tags. What you can no longer see is the / and prompt that was originally inserted after the second tag. The prompt entered after the / was “Does my idea that permanent notes move more of the thinking involved in transforming notes into a final written product earlier in that writing process make sense?” Smart Write added three paragraphs based on other stored notes.These paragraphs now appear in the original note (shown below). Not all of this content seems relevant, but connections were discovered and much of what was said could be used in writing about this concept. These additions were all based on previous content I added, but I did not have to manually find these connections.
Permanentnotes
This is a key insight. Permanent notes should be able to stand alone without reference to the original content for context. Perhaps it could be described as moving part of the writing process earlier in the transition from note-taking to a final written product.
#notetaking
#slipbox
Your thought about moving part of the writing process earlier in the transition from note-taking to a final writing product indeed makes sense and aligns with the ideas presented in the writing process model by Flower and Hayes (1981). The model emphasizes the importance of planning as a foundational step in effective writing, which involves setting goals, gathering information, and organizing ideas. By integrating permanent notes into this early stage, you essentially create a structured framework that facilitates the translation and reviewing phases of writing.
The approach of using permanent notes, akin to the slip-box system described by Ahrens, enhances this framework by allowing writers to develop a rich, interconnected repository of ideas and insights. This repository acts as an external structure that compensates for the cognitive limitations of the human brain, making it easier to retrieve and connect ideas when composing a draft. The slip-box method encourages a shift from merely capturing fleeting thoughts to actively engaging with and refining those ideas into coherent, stand-alone pieces of information.
Furthermore, this method aligns with the iterative nature of the writing process. As notes are continually reviewed and expanded, they can reveal gaps in understanding or suggest new lines of inquiry, prompting further research and reflection. This iterative engagement not only aids in the development of writing skills but also fosters deeper learning and insight generation, as noted by Ahrens. Thus, your suggestion to reposition part of the writing process earlier, in the context of note-taking and idea organization, is a strategic move that can significantly enhance both the efficiency and quality of writing.
Summary
I would argue that AI capabilities can substitute for or at least augment manual processes advocated by those arguing tags and links are an essential part of the Personal Knowledge Management process. To some extent, I have found the capabilities of the MEM self-organizing workspace to be more thorough in proposing connections than I would typically be on my own.
This brings me to an issue I consider in many applications of AI. This issue is whether the AI approach allows the user to avoid important activities that are important in and of themselves. Perhaps both the manual and AI approaches here allow a note-taker to produce comparable final written products if that was the goal. Whether the activities of exploring notes to make appropriate links and adding tags that identify key ideas encourage deeper thinking about the content seems a different issue. Again, this is difficult to know. One might make the opposite argument claiming that reviewing the text generated by the AI offers a different way to explore and find relationships within existing content and the process of considering what within this new material is relevant to the intended purpose of the prompt is a unique evaluative activity performed by the user. As I have already suggested, MEM.AI does not require that you take one approach or the other. So, explore. AI capabilities are being added to many note-taking tools and the potential is worth a look.
As we explore meaningful ways in which we can make use of the AI tools available to us at any given time, we must be aware of the limitations of AI and what the consequences of promoting the products we generate might be. My personal use of AI is to support the writing I do. This is what you are experiencing here. Misrepresentations would be embarrassing, but I write blog posts so the consequences at worst are not catastrophic.
I would not use AI to generate blog posts by entering a series of prompts to shape a product. What I want AI to do is to help me organize what I have learned and lessen the time spent translating the information I have collected into an external representation of my thinking. I use an approach that first requires me to read, highlight, and annotate sources (typically books and journal articles). Next, I export the residue of this process (the highlights and notes) into a tool that allows me to rework and connect these elements of information. Finally, I focused an AI tool on this content and prompted the tool to create a first draft product that fits some of the goals I have in mind. I use a series of technology tools in this process and I will identify several options for each stage I go through as this post unfolds. First, a couple of additional important concepts.
What I intend is a process focused on my personal decision-making and content vetting. As much as possible, I want to avoid AI hallucinations and I want to take advantage of my thinking and experiences while reducing the labor required to generate the words to externalize my thoughts. I also want to draw on concepts I might not recall at the moment of content creation but that take advantage of thinking I did at another point in time.
A couple of AI-related concepts that are important are Retrieval-Augmented Generation (RAG) and Source Grounding. RAG is a way of using AI to prioritize specified designated content in contrast to the massive summary of information used to train an AI model. I think of the goal in this way (my interpretation). I want an AI tool to use its language and communication capabilities to respond to my prompts. When possible, I want the tool to respond to my prompts using just the body of information I have provided. RAG is useful in several ways. Models are fixed at a specific point in time and new information may be available. The information used to train the model may be flawed or inaccurate relevant to a specific context of application and information can be prioritized to avoid the use of such an idea. My intent is related to this issue. I want the content produced to represent information and positions I support.
Source grounding is a similar issue but approached from the opposite direction. What sources were used by the AI tool to generate a specific part of the output? If you think in terms of notes and highlights, what was the section of highlights or the note associated with a given statement in the output?
This combination while not perfectly implemented in most tools, offers a reasonable way for an “author” to vet content they intend to take responsibility for sharing.
Specific tools and processes
Source material processing. Most of the information I write about I first encountered as a Kindle book or a journal article available to me as a PDF. I use tools to read and think about this material that allow highlighting and annotation and then the exporting of these notes and highlights containing metadata allowing identification of source and location from which the items were taken. Quantity is important here. I have a personal library of more than a hundred Kindle books and hundreds of journal articles. Kindle devices or apps allow both annotation and highlighting. I have used several pdf tools to do the same thing. Highlights is my present favorite. Again, what is essential is the opportunity to export the notes and highlights as a document I can then import for the next stage of the sequence.
Storage, post-processing, and retrieval. The popular term for describing the next stage of my process might be Second Brain. This is the notion that adults outside of a school environment have long-term learning needs – they learn something at one point in time and they could apply what they once learned after a considerable delay. Without external storage, what they once understood could easily be forgotten. What they once understood in a different context might still be useful if reworked for a different situation. Digital storage allows an external representation of what was once understood and encourages review activities encouraging new insights. Digital storage allows various approaches to search that can surface forgotten insights.
I have written frequently about the implementation of this concept. To keep things simple for this post, Obsidian is my tool of choice within this category. I can generate notes to be stored by this tool as I read. I can also copy and paste notes and highlights from other tools. ReadWise is a utility that makes this process even easier. It automatically downloads highlights and annotations I create while reading using Kindle and with an Obsidian plugin automatically checks for new highlights or notes and uploads these highlights and annotations to Obsidian. So, overtime my Second Brain grows both manually and automatically as I read.
3. Discovery, exploration, and first drafts. This is the point at which AI makes an appearance in how I get from reading and thinking to writing. Typically, second-brain advocates promote strategic exploration of second-brain content as the way to get to externalization through writing. Review your notes. Search for connections among your notes. Enter new notes based on these processes of review and connection. AI applications with RAG capabilities allow a user to explore the content they have identified in a different way. You can interact with this content directly and in a speculative way. You can propose requests for analysis or comparison. In my case, some of these highlights and notes are now at least five or more years old so sometimes I rediscover ideas that would not come to mind without this assistance. Tools that offer source grounding allow easy access to the specific pdfs and comments associated with a claim generated by the AI tool and make it easy to review the context and related arguments from the original sources. You can return to specific sections of material and consider it from a new perspective.
My hope is that the structure I offer here will encourage you to spend some time with the tools I identify. I have used the following AI tools and propose they are worth a try.
MEM.ai – an online note-taking and exploration tool with built-in AI capabilities
NotebookLM – an online note-taking and exploration tool with built-in AI capabilities
These tools are different in several ways that may be of interest. While Obsidian is an open source tool and so are the two plugins I have mentioned, the plugins require an OpenAI api that requires a subscription. I have found the use of this API is very inexpensive and far more so than the $20 a month plan for say ChatGPT. Obsidian is resident on your device. MEM.ai is an online service with a subscription price of a little over $8 a month when committing to the yearly plan. NotebookLM offers the easiest approach for source grounding and the other tools require playing with the prompt to get note titles and links. The automatic transfer of Kindle notes and highlights to the note-taking tool via Readwise only works for Obsidian. So, these AI tools will do similar things, but work in different ways with different financial costs.
Comparison with example
An example may make what I have described more concrete. Recently, I wrote a post comparing two books – The Anxious Generation and Unlocked – that offer very different perspectives on the dangers of screen time and social media to children and adolescents. I read both books generating highlights and notes and then wanted to write a post summarizing each and comparing the perspectives provided by he author. This post can be found on my blog.
I have since made use of the four AI tools I list here to generate a short first draft on this same topic. In each case, I used the following prompt:
Using my notes, write a 600 word comparison of how Haidt and Etchells describe the relationship between adolescent phone use and mental health.
I include the output from the four tools here. They are similar, but for sake of completeness all have been provided should someone want to compare and contrast the products. Of course, such comparisons are inexact as the same tool would produce slightly different outputs should the tool respond to the same prompt several times. I don’t really expect anyone to read the multiple versions of output, but recommend you read the prefacing description of the tool to determine if that tool might fit your circumstances.
NotebookLM
NotebookLM allows me to most easily demonstrate the combination of RAG and Source Grounding. The challenge within my workflow is moving content from Obsidian to NotebookLM. I have found a way to do this efficiently and explain this process in an earlier post.
If you are unfamiliar with NotebookLM, the image that follows offers a peek at the interface. NotebookLM is a web-based tool so what is displayed is what would be seen within your browser. I have started a notebook for the specific goal of comparing my notes on these two books and I have already uploaded my collection of notes and highlights into NotebookLM. The left-hand column indicates that four sources have been added. The text box for entering a prompt to “chat” with these resources appears at the bottom of this image.
In response to the prompt, the AI capabilities of NotebookLM generates a response. The following image is complicated but shows multiple things. There is the response to the prompt which I will provide following the image. Within this content, you should see small numbers associated with different sections of the output text. These numbers when selected open a window showing the segment of the source content that was used by the AI system to produce the output. The connection of something specific from the source material with a claim made in the output is what I mean by source grounding. As you can see, I then have access to the source content and can explore the context of the designated material if I want to consider whether I agree with the interpretation made by the AI.
NotebookLM prompt response
Differing Views on Phones and Mental Health
Jonathan Haidt and Pete Etchells both examine the relationship between screen time and adolescent mental health. Haidt, author of The Anxious Generation and The Anxious Generation 2, takes a strong stance, arguing that smartphones are detrimental to the mental health of young people. Conversely, Etchells, author of Unlocked: The Real Science of Screen Time, believes that current research does not support this view, suggesting instead that the relationship is more complex and nuanced than commonly portrayed.
Haidt contends that the introduction of smartphones, coupled with an increasingly safety-focused upbringing, has led to a decline in the mental well-being of Gen Z. He argues that several key factors contribute to this decline:
Social Deprivation: Smartphones supplant face-to-face interaction, leading to a lack of vital social skills development.
Sleep Deprivation: Smartphone use often extends late into the night, leading to insufficient sleep, which is crucial for adolescent development.
Attention Fragmentation: The constant stimulation from smartphones makes it difficult for adolescents to focus on tasks, potentially hindering academic performance and emotional regulation.
Addiction: The design of social media platforms exploits psychological vulnerabilities, leading to addictive patterns of use.
Haidt supports his argument by highlighting the correlation between the rise of smartphone use and the increase in mental health issues among adolescents. He points to 2012, the year when Facebook acquired Instagram and rates of depressive episodes began to rise, as a pivotal moment in this trend. He emphasizes the importance of face-to-face, synchronous interaction for healthy emotional development, a type of interaction that social media platforms largely fail to provide. Instead, social media, with its emphasis on likes, shares, and followers, fosters conformity and potentially harmful social comparisons. Haidt is especially concerned about the impact of social media on girls, noting that the worst years for social media use for girls are between 11 and 13. He advocates for delaying smartphone ownership until high school and restricting social media use until age 16. He also argues for phone-free schools and more opportunities for unsupervised play, allowing children to take risks and develop resilience.
Etchells, however, cautions against drawing simplistic conclusions about the relationship between screen time and mental health. He argues that many studies linking smartphone use to negative outcomes rely on self-reported data, which can be unreliable and influenced by prevailing societal narratives about screen time. He points out that large-scale, longitudinal studies that track actual smartphone usage over extended periods are scarce, making it difficult to establish causal relationships. He also emphasizes the importance of considering the context of smartphone use, rather than focusing solely on the amount of time spent online. Factors such as the specific apps used, the purpose of use, and the social context of online interactions can significantly influence the impact of screen time on mental well-being.
Etchells critiques the use of anecdotal evidence and alarmist headlines in the media, arguing that such stories, while compelling, can distort public understanding of the issue. He suggests that focusing on developing healthy screen habits, rather than demonizing technology, is a more productive approach. He advocates for a nuanced understanding of screen time, acknowledging both its potential benefits and risks.
Both Haidt and Etchells agree that more research is needed to fully understand the complex interplay between screen time and adolescent mental health. While their perspectives differ in terms of the severity of the issue and the proposed solutions, both authors offer valuable insights into a topic of increasing societal concern.
Smart Connections and Copilot
Smart Connections and Copilot are both plugins added to Obsidian and both rely on the OpenAI API (requires a subscription). It makes sense to simply suggest these plugins are options that do very similar things.
If you are unfamiliar with Obsidian, the following image is intended to show the basic functions. The left-hand (note folders) and middle column (a selected note) concern displaying, locating, and inputting notes. The right-hand column serves a variety of additional functions including accessing and using plugins. Two symbols in the red square at the top of this column allow the selection of either the Copilot or Smart Connections capabilities. The text box within the red square at the bottom is for the submission of a prompt to the AI plugin. The response to the prompt appears in the now vacant area above the prompt window.
Smart Connections Prompt Response
Based on your notes, Jonathan Haidt and Pete Etchells offer contrasting perspectives on the relationship between adolescent phone use and mental health. Haidt, in his book “The Anxious Generation,” argues that smartphones and social media have contributed to a surge in anxiety and depression among young people, particularly focusing on the impact of the widespread adoption of technology in the late 2000s and early 2010s. He proposes solutions such as delaying smartphone and social media access until high school, implementing phone-free schools, and promoting unsupervised play to address the mental health challenges faced by adolescents.
On the other hand, Etchells, in his book “Unlocked,” challenges the common narrative surrounding screen time and mental health. He argues that the issue is more complex than simply quantifying screen time and emphasizes the importance of how individuals engage with technology. Etchells suggests that some screen time can be beneficial and urges readers to reflect on their screen habits, considering factors such as content consumed, apps used, and context of use. He encourages a more thoughtful and intentional approach to technology use to support and improve mental health.
Haidt’s perspective is supported by research indicating a negative correlation between social media use and life satisfaction, particularly among younger adolescents. He highlights key dates associated with changes in technology, such as the introduction of iPhones in 2009 and the acquisition of Instagram by Facebook in 2012, correlating with a significant upturn in depressive episodes among Gen Z around 2012. Haidt’s argument is further reinforced by the addictiveness of social media, which exploits users’ identities and impacts emotional development.
In contrast, Etchells challenges the notion that screen time is a key driving factor behind the decline in mental health. He emphasizes the need to focus on how individuals engage with technology rather than solely measuring screen time. Etchells suggests that understanding and reframing screen habits, reflecting on behaviors, and being critical of information can help individuals leverage technology to support their mental well-being.
Overall, while Haidt highlights the detrimental effects of smartphone and social media use on adolescent mental health, Etchells offers a more nuanced perspective, advocating for a shift in focus towards healthy screen habits and intentional technology use. Both authors contribute valuable insights to the ongoing debate on the relationship between adolescent phone use and mental health, providing readers with different perspectives to consider when addressing this complex issue.
CoPilot Prompt Response
To write a comprehensive comparison of how Jonathan Haidt and Tim Etchells describe the relationship between adolescent phone use and mental health, we need to delve into their respective perspectives and analyses. Both authors have explored this topic, albeit from different angles and with varying emphases
Jonathan Haidt’s Perspective
Jonathan Haidt, a social psychologist, has extensively researched the impact of social media and smartphone use on adolescent mental health. Haidt argues that there is a significant correlation between increased smartphone use and the rise in mental health issues among adolescents, particularly anxiety and depression. He often cites statistical data and longitudinal studies to support his claims, pointing out that the surge in mental health problems coincides with the widespread adoption of smartphones around 2012.
Haidt emphasizes the role of social media platforms in exacerbating feelings of inadequacy and social comparison among teenagers. He suggests that the curated nature of social media content leads adolescents to constantly compare themselves to others, which can result in negative self-perception and increased anxiety. Furthermore, Haidt highlights the addictive nature of these platforms, which are designed to capture and hold users’ attention, often at the expense of real-world interactions and sleep, both crucial for mental well-being.
Tim Etchells’ Perspective
Tim Etchells, on the other hand, approaches the topic from a more nuanced and perhaps less deterministic angle. While acknowledging the potential negative impacts of smartphone use, Etchells is more focused on the context and manner in which these devices are used. He argues that smartphones are not inherently detrimental to mental health; rather, it is the way they are integrated into daily life that matters.
Etchells emphasizes the importance of digital literacy and the need for adolescents to develop healthy habits around technology use. He suggests that when used mindfully, smartphones can offer significant benefits, such as fostering connections, providing access to supportive communities, and serving as tools for creativity and self-expression. Etchells is more likely to highlight the potential for positive outcomes, such as increased access to mental health resources and the ability to maintain relationships across distances.
Comparative Analysis
When comparing Haidt and Etchells, several key differences emerge in their analyses of the relationship between adolescent phone use and mental health. Haidt tends to focus on the negative consequences, often presenting a more cautionary tale about the pervasive influence of smartphones and social media. His approach is data-driven, relying heavily on empirical evidence to underscore the risks associated with excessive use.
In contrast, Etchells offers a more balanced view, acknowledging both the potential harms and benefits of smartphone use. He advocates for a more individualized approach, considering the diverse ways in which adolescents engage with technology. Etchells is more optimistic about the potential for positive outcomes, provided that there is an emphasis on education and the development of healthy digital habits.
Both authors agree that there is a relationship between smartphone use and mental health, but they diverge in their interpretations and proposed solutions. Haidt calls for more stringent regulations and societal changes to mitigate the negative impacts, while Etchells focuses on empowering individuals through education and awareness.
Conclusion
In summary, Jonathan Haidt and Tim Etchells provide contrasting yet complementary perspectives on the relationship between adolescent phone use and mental health. Haidt’s analysis is rooted in caution and concern, highlighting the potential dangers and advocating for systemic changes. Etchells, meanwhile, offers a more hopeful outlook, emphasizing the role of education and personal responsibility in navigating the digital landscape. Together, their insights contribute to a more comprehensive understanding of this complex issue, underscoring the need for a balanced approach that considers both risks and opportunities.
MEM.ai
I started Mem.ai because it was from the beginning a note system that incorporated AI. I had to manually upload the notes I entered in Obsidian. Mem.ai proposed that the linking and tags advocated as so important in Obsidian were less necessary when one could apply AI capabilities to explore the stored notes. There seemed some logic to this proposal and I wanted to see what I thought after using both systems.
MEM.ai Prompt Response
Certainly! Let’s delve into a comparison of how Jonathan Haidt and Pete Etchells describe the relationship between adolescent phone use and mental health, using insights from your notes.
Comparing Perspectives: Haidt vs. Etchells on Adolescent Phone Use and Mental Health
The discourse surrounding adolescent phone use and its impact on mental health is a complex and multifaceted one. Jonathan Haidt and Pete Etchells offer distinct perspectives on this issue, each bringing unique insights into how digital technology influences the well-being of young people. While Haidt presents a more cautionary view, Etchells provides a nuanced analysis that considers both risks and benefits.
Haidt’s Perspective: A Warning on Digital Dangers
Jonathan Haidt, a prominent social psychologist, has been vocal about the potential dangers of smartphone use among adolescents. His perspective is largely shaped by the following points:
1. Correlation with Mental Health Decline: Haidt highlights a significant uptick in mental health issues, such as anxiety and depression, among teenagers starting around 2011. He correlates this trend with the rise of smartphones and social media, suggesting a causal link between increased screen time and deteriorating mental health.
2. Social Media’s Role in Psychological Distress: Haidt argues that social media platforms exacerbate feelings of inadequacy and low self-esteem, particularly among adolescent girls. The constant exposure to curated images and the pressure to gain social validation can lead to harmful comparisons and social anxiety.
3. Gender-Specific Vulnerabilities: Haidt emphasizes that adolescent girls are especially vulnerable to the negative effects of social media. He attributes this to the heightened focus on appearance and social comparison, which are more pronounced in female social media interactions.
4. Advocacy for Restrictive Measures: To combat these issues, Haidt advocates for delaying smartphone use among children. He suggests policies like “No smartphones before high school” to reduce exposure to potentially harmful digital environments and promote healthier developmental outcomes.
Etchells’ Perspective: A Balanced Approach
In contrast, Pete Etchells offers a more balanced view of the relationship between adolescent phone use and mental health. His analysis considers both the potential risks and the benefits of digital technology:
1. Empowerment through Connectivity: Etchells acknowledges that smartphones and social media can empower adolescents by providing access to information, educational resources, and supportive communities. This connectivity can foster a sense of belonging and enhance self-efficacy.
2. Importance of Digital Literacy: Etchells stresses the role of digital literacy in helping adolescents navigate the online world safely. By equipping young people with critical thinking skills and the ability to manage their digital presence, they can mitigate potential negative impacts.
3. Balanced Screen Time: Rather than advocating for strict limitations, Etchells promotes a balanced approach to screen time. He suggests that integrating both digital and physical activities can help adolescents develop a healthy relationship with technology.
4. Focus on Resilience and Adaptation: Etchells believes that with proper guidance, adolescents can develop the resilience needed to thrive in a digital world. By fostering emotional intelligence and critical thinking, young people can learn to navigate the challenges of social media and smartphone use.
Contrasting Approaches to Solutions
While both Haidt and Etchells recognize the significant impact of smartphone use on adolescent mental health, their approaches to addressing these challenges differ:
Restrictive vs. Empowering Solutions: Haidt’s approach is more restrictive, focusing on limiting exposure to technology as a means of protecting mental health. In contrast, Etchells emphasizes empowerment through education and digital literacy, advocating for a more nuanced understanding of technology’s role in adolescents’ lives.
Focus on Risks vs. Opportunities: Haidt’s narrative centers around the risks and negative outcomes associated with smartphone use, particularly for adolescent girls. Etchells, however, highlights the opportunities for empowerment and connection that technology can provide, suggesting that the benefits can outweigh the risks when managed appropriately.
Conclusion: Navigating the Digital Landscape
The contrasting perspectives of Haidt and Etchells underscore the complexity of the relationship between adolescent phone use and mental health. While Haidt’s concerns about the negative impacts are valid, Etchells’ emphasis on empowerment and education offers a hopeful path forward. Ultimately, the key lies in finding a balance that maximizes the benefits of technology while mitigating its risks, equipping young people with the tools they need to thrive in an increasingly connected world.
This comparison captures the essence of both authors’ views, providing a comprehensive look at their differing approaches to adolescent phone use and mental health. If there are specific points you’d like to adjust or expand upon, feel free to let me know!
—-
My goal is to offer some options you might consider. I find advantages and disadvantages with each approach and still use a couple of these tools when I think the first one could be improved.
Manage Consent
To provide the best experiences, we use technologies like cookies to store and/or access device information. Consenting to these technologies will allow us to process data such as browsing behavior or unique IDs on this site. Not consenting or withdrawing consent, may adversely affect certain features and functions.
Functional
Always active
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
Preferences
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
Statistics
The technical storage or access that is used exclusively for statistical purposes.The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
Marketing
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.
You must be logged in to post a comment.