AI tutoring now


You can make cognition as simple or complex as you want, but for me, learning boils down to thinking about experiences and being an effective educator boils down to encouraging and perhaps directing this thinking in students.

Research conducted by educational and cognitive psychologists provides some ideas about how this directing and encouraging might be accomplished. I always emphasize the “might” in my writing. Simply put, learning must be accomplished by the learner and another person or an external task can only work when a learner engages with that person or that task.

I find the research cognitive psychologists conduct incredibly useful in understanding what it takes to accomplish learning and based on this understanding how educators and learners might then proceed to be more successful. Just as every gardener does not have to study botany and chemistry to achieve a great harvest of tomatoes, not every researcher must immerse herself or himself in the details and depth of research to be a successful educational practitioner. Some practitioners may be fascinated with how understanding and retention are accomplished and others just may want to know what tasks to assign that are usually productive. As an academic retiree, I  spend some time as a translator between these two types of individuals. I would also propose that the interaction of these two groups is important because ideas flow in both directions. For science-based approaches to change and improve both groups should be willing to share their perspectives and be thoughtful in interpreting the ideas presented to them.

A couple of years ago I became interested in a group of online services that were developed to improve the educational value of online content (web pages and videos). I developed my own way of describing what these services were developed to accomplish. It seemed that these services were designed to transform informational content into learning experiences and this was accomplished by layering different elements or tasks on the original content. Layering seemed the correct description because these services could not actually modify the content originally shared by content creators for ethical and legal reasons, but what the service could do was take the feed from the creator’s service and add elements on top. Elements were additions that could encourage important cognitive behaviors in a learner. I typically describe these elements as external tasks that encourage cognitive or mental behaviors. This description is similar to what others call generative activities (to be accurate generative activities are sometimes described as requiring the creation of a product, but the idea of an external task to influence cognitive behavior is the same).  Questions are my favorite example of an external activity that can be added to encourage a variety of important thinking (internal) behaviors. For example, if you want a learner to link a new concept to everyday experiences the concept is useful in understanding, you might ask the learner to provide examples that show the application of the concept. Many learners may do this without the question, but the question increases the likelihood more learners will work to identify such connections with their existing experiences. Those who think about instruction in this way may describe what they are doing as designing instruction. What are the cognitive tasks that are important in learning a certain thing and what external activities might be added to the exposure to the input of demonstrations and other inputs to increase the probability these cognitive behaviors occur?

For example, what external tasks would increase the likelihood of:

  • Knowledge activation – what do I already know that is relevant
  • Elaboration – how does this input fit with what I know and what might it imply
  • Comprehension monitoring – do I understand, does this apply as I think it does
  • Retrieval practice – what makes this easier to retrieve when it is appropriate

Depending on the specific service, the elements that layering services provided included annotations, highlighting, questions, and discussion prompts. Annotations could include additional material such as examples, translations, or instructions. Questions could be open-ended or multiple-choice. A few of these elements could also be added by the learner (highlights and annotations) so elements provided to the designer could be used to encourage specific use of the elements available to students.

The application of this approach to AI-assisted study behavior

My interests seem to occur in waves. There was the layering wave and I, like so many others, have launched into an AI in education wave. AI has rapidly taken over most of the online discussions I follow. AI can be applied in many ways, but I think at this time one of the most productive is study behavior. One of my core interests has always been study behavior. I define study behavior broadly as everything that follows the initial exposure to new experiences. Note-taking in class occurs after the initial exposure. The questions a teacher asks or the discussions that are encouraged with peers during class occur after the initial exposure. Then, there are all of the activities, required and voluntary, students engage in on their own. For those of us outside of a formal educational setting, we may engage in such activities as taking notes or highlighting what we read to enhance the benefit of pure exposure to new material. 

My new interest is in how AI tools can be applied as a tutor or skilled peer to improve the study experience. There is a good deal of overlap between the approach I have been taking with promoting layering services and now with AI tutoring. Tutoring is possibly the best way to assist a learner, but tutoring is expensive and access is not always available. If we consider what a tutor does for a learner and also take the approach I describe as identify productive external tasks that promote learning, we may find similar opportunities when treating some AI services as a tutor.

What I am proposing here is not that novel and efforts to create smart tutors have generated commercial and research interest for years and presently by organizations with connections to education (Google, Khan Academy). What I think I can add are some ideas about using an available AI tool (ChatPDF) right now to explore this capacity. 

ChatPDF as the name implies allows an AI-assisted interaction with a PDF that you upload. Focusing AI on a specifically designated resource has both advantages and disadvantages. With studying, you are often studying something specific – a specific resource. The information you are encountering might be available from many other sources, but if you are a student, your responsibility can be defined in terms of specific sources. The advantage I have found with using a specific resource rather than “general knowledge” with present AI applications is that my interaction seems more accurate with a target source. I have less challenges with an AI chat service wandering off and offering content that is identifiably erroneous. This will likely change as improvements continue, but for now there are problems. The disadvantage of targeting a specific source is that aspects of studying involve going beyond the information provided to speculate about examples and applications not presented in the target material. When you constrain the interaction to a source, I have found AI struggles with this type of expectation. AI tools now available can do many impressive things, but you do need to have a sense of what is reasonable to expect and what problems might occur. 

I recognize that many resources a student encounters are not provided as pdfs (e.g., a textbook). I have no immediate workaround for this challenge so I offer this post as a recommendation for exploration. I predict this is what is coming and that there are opportunities available now in specific situations. For my own exploration, I have been using the content I wrote. I want to make certain I understand the content I am applying AI to investigate at a deep level so I can evaluate what the AI offers me as a learner.  I have authored a couple of textbooks and I was able to take the word processing file of a chapter, save it as a pdf, and then upload it to ChatPDF (here is the chapter if you want to submit it yourself and mimic my prompts). 

As I explored ChatPDF and thinking about arguing the service could act as a tutor, I decided that that comparison to a professional tutor might be a bit much. Perhaps, more accurately, the AI service could be described as acting like a capable peer. A talented peer would be helpful and accurate most of the time, but you might also expect that sometimes you would find things your peer did not quite get as well as you did and still find it helpful to continue to study with this individual. 

What follows are some of my interactions with the chapter I uploaded using ChatPDF. I label the material in the following may – Comment, Prompt, and AI – so you can interpret the various examples I provide. Remember that if the AI presents a question and the learner responds, the question is the AI presenting and the answer is the learner returning a Prompt. Prompts are what the learner submits and AI is always what the service replies in the material that follows. The idea with these examples is to offer some insights into how this tool could be used by a learner to assist them in improving their understanding and retention of content. Unlike a human tutor, a learner must prompt the AI system to create the interaction. However, the system is quite versatile and it is worth the time to just explore and see what it will do. As you will see, the system allows a conversation to continue through several interactions so there may be multiple prompts and responses that continue on the same initial theme. This would be the way a student and tutor interact.  

ChatPDF Examples

[Comment] Gagne’s 9 events, a very popular model of instructional design, suggests a learning experience begin with an event intended to capture learner attention. To me, using ChatPDF after reading assigned content is a study experience that would not typically include this element, but just to offer an example of how an AI service might generate such an event see the following.  

[Prompt] Identifying something from this chapter that will likely surprise me.

[AI] Something surprising and interesting idea from this chapter is the concept of “personal theories,” which are individual beliefs or mental models that people develop based on their experiences and observations. These personal theories can sometimes be flawed or inaccurate, and may lead to inert knowledge if they are not challenged or corrected by more formal, accurate theories learned in an educational setting. This highlights the importance of not just teaching information, but also helping students to develop critical thinking skills and the ability to evaluate their own beliefs and assumptions.

[Comment] Activate existing knowledge. Events two and three in Gagne’s model are to identify goals and activate relevant existing knowledge. Again, I am focusing here on how a tutor might guide study, but again I will show a use of the AI tutor that accomplish such goals and with this example I will show how a learner could followup as one might with a human tutor. 

[Prompt] What are the most important ideas I should learn from this chapter?

[AI] This chapter covers several key ideas related to the nature of learning and the thinking processes necessary to produce learning. Some of the most important ideas include understanding how students learn, what skills and knowledge they should acquire, and how to design effective learning tasks that promote meaningful learning. Additionally, we discuss the cognitive system at both a detailed and conceptual level, identifying core functions that contribute to student learning. Finally, we emphasize the importance of considering the limitations of the cognitive system when designing instruction.

[Prompt] What are questions I should be able to answer to show I have learned these important ideas?

[AI] To demonstrate your understanding of the important ideas covered in this chapter, you might be able to answer questions such as:

  • What are some key factors that influence how students learn?
  • How can you design learning tasks that promote meaningful learning?
  • What are some core functions of the cognitive system that contribute to student learning?
  • Why is it important to consider the limitations of the cognitive system when designing instruction?
  • How can you use your understanding of how students learn and what skills and knowledge they should acquire to inform your teaching practice?

[Prompt] Let me pick one of these questions to see if I understand.

[AI] Sure, feel free to pick any of the questions I provided earlier and I’ll do my best to help you answer it!

[Prompt] What are some limitations of the cognitive system that educators should consider when designing instruction?

Learners can easily become overloaded when learning something new. Their existing knowledge may also not provide necessary background knowledge and they need time to process new ideas. If more and more information is added before students have time to think, they will miss things when they become overloaded.

[AI] You did a good job summarizing the idea that learners can become overloaded when presented with too much information at once. It is important for educators to consider the limitations of the cognitive system when designing instruction, such as the limited capacity of working memory and the importance of providing sufficient time for learners to process new information. By taking these limitations into account, educators can design instruction that is more effective and helps learners avoid becoming overloaded.

[Comments] One of the easiest uses of AI is to use it to ask and respond to questions. You can test your understanding of a claim you make or ask the service to present you with questions. The wording of the prompts is important in getting the service to interact with you in the way you want. Read the way I ask the service to present me with a series of multiple-choice questions carefully.

[Prompt] Is the limit of short term memory what I can remember the next day

[AI] According to page 3 of this PDF, there is a limit to how long information will be maintained in working memory without continued attention. This means that the limit of short-term memory is not necessarily what you can remember the next day, but rather how much information you can hold in your mind at any one time without actively rehearsing or attending to it.

[Prompt] If I am a teacher, how can I provide a form of metacognition for a student when that student cannot provide it for himself?

[AI] As a teacher, you can provide metacognitive support to students who may not be able to provide it for themselves. One way to do this is by modeling metacognitive strategies and explicitly teaching students how to use them. For example, you might demonstrate how to monitor comprehension while reading by thinking aloud and sharing your thought process with the class. You could also provide prompts or questions that encourage students to reflect on their own learning, such as “What did you learn from this activity?” or “What strategies did you use to solve this problem?” Additionally, you could encourage peer-to-peer discussions where students can share their own insights and perspectives on the learning process. By providing these types of supports, you can help students develop their metacognitive skills and become more effective learners.

[Prompt] Ask me a short essay question based on the content of this chapter.

[AI] How can teachers help students to develop critical thinking skills and avoid the phenomenon of “inert knowledge”? Provide specific examples of instructional strategies that can be used in the classroom to promote active learning and meaningful application of new information.

[Prompt] Remind students they learned information relevant to a novel problem in class as part of challenging them with a new problem to solve.

[AI] This statement aligns with the concept of “transfer” discussed in the chapter on “Inert Knowledge.” Transfer refers to the ability to apply knowledge and skills learned in one context to new situations or problems. One way that teachers can promote transfer is by explicitly connecting new problems or tasks to previously learned information or skills. By reminding students of what they have already learned and how it relates to the current problem, teachers can help students make connections and apply their knowledge in a more meaningful way. This can also help to avoid the phenomenon of inert knowledge, where students may have acquired information but are unable to use it effectively in new situations.

[Comment] A limitation is revealed in the following interaction. The tutor asks a question requesting an example and the example given is not something the tutor understands to be a part of the chapter. 

[Prompt] Ask me a short essay question based on this chapter

[AI]] How can the concept of “inert knowledge” impact student learning, and what strategies can teachers use to help students avoid this phenomenon? Provide specific examples to support your answer.

[Prompt] A parent knows that young children may commit some bad behaviors to get his parents attention, but typically parents respond to such misbehaviors anyway. 

[AI] I’m sorry, but I’m not sure how this statement relates to the chapter on “Inert Knowledge” and learning strategies. Is there a specific question or topic you would like me to address?

[Prompt] Generate 10 multiple choice questions for this document. Number the questions and allow me to provide an answer for each number. Do not provide answers I can see.

[AI] Sure, here are 10 multiple choice questions based on the document:

What is metacognition?

a) The capacity to direct and control our own cognitive behavior

b) The ability to memorize information quickly

c) The process of learning through repetition

d) The use of external tasks to support weak cognition

What is the common challenge when working with a group of learners?

a) Meeting individual needs when working with a group

b) Encouraging students to work independently

c) Providing enough resources for all students

d) Ensuring that all students have the same level of knowledge

What is an example of metacognitive control?

a) Monitoring your own understanding of the material

b) Memorizing information without understanding it

c) Repeating information until it is learned

d) Using external tasks to support cognition

What is metacognitive knowledge?

a) Personal insights regarding the self and tasks and potential strategy options

b) Memorized information about a subject area

c) The ability to repeat information quickly and accurately 

d) The use of external tasks to support cognition

How can teachers provide metacognitive support to students?

a) By modeling metacognitive strategies and explicitly teaching students how to use them 

b) By providing lectures without any interaction from students 

c) By assigning homework without any guidance or feedback 

d) By providing multiple-choice tests only 

What is an example of an external task that may support weak metacognition?

a) Study questions 

b) Group discussions 

c) Peer-to-peer feedback 

d) Individual reflection 

What does effective tool use require?

a) Recognizing task demands, personal strengths and weaknesses, and insight necessary to recognize and respond when a plan does not achieve expected results.

b) Memorizing information quickly without understanding it.

c ) Repeating information until it is learned.

d) Using external tasks to support cognition.

What is the capacity to direct and control our own cognitive behavior?

a) Metacognition 

b) External tasks 

c) Cognitive processing 

d) Learning strategies 

What is the importance of understanding how students learn?

a) It helps teachers design effective learning experiences for students

b) It helps teachers memorize information quickly

c) It helps teachers assign homework without any guidance or feedback

d) It helps teachers provide multiple-choice tests only

[Prompt] 1 a

[AI] Correct! Metacognition is the capacity to direct and control our own cognitive behavior.

[Prompt] 2 a

[AI] Correct! The common challenge when working with a group of learners is meeting individual needs when working with a group.

etc.

[Comment] Here is a sequence generated when I asked for help understanding, but was unclear about whether the AI bot “understands”  a concept in the same way I use the concept. For example, I understand that inert has to do with what context triggers the use of knowledge. Hence, a student who understands the meaning of a word on a test or when asked in class would not necessarily access the same knowledge when in a different context (at home reading a book) that does not trigger “school content”. This seems how I used the term when writing the cognitive chapter. I ended up trying to negotiate an interpretation by interacting with AI as tutor which while not generating a simple clarification I might have wanted was fairly sophisticated. I found myself arguing with the AI service about the meaning of something I had written, but this interaction was interesting. 

[Prompt] Help me understand inert learning. What does the term mean and provide an example.

[AI] Inert learning refers to the phenomenon where students may learn information but are unable to apply it in new or unfamiliar contexts. In other words, they may be able to repeat information but are not able to use it in a meaningful way. An example of inert learning might be a student who memorizes a list of vocabulary words for a test but is unable to use those words in conversation or writing outside of the test context. This can happen when students are taught information in isolation, without opportunities to practice applying it in different contexts or with different types of problems. To avoid inert learning, it is important for teachers to provide opportunities for students to practice using new information in different ways and contexts.

[Prompt] What do you mean by test context? What do you mean by a meaningful way?

[AI] By “test context,” I mean the specific situation in which the student is asked to demonstrate their knowledge, such as on a quiz or exam. In this context, the student may be able to recall information that they have memorized, but they may not be able to apply it in a more complex or unfamiliar situation.

By “meaningful way,” I mean using the information in a way that is relevant and useful to the student’s life or goals. For example, if a student is learning about fractions, being able to use that knowledge to solve real-world problems involving measurements or cooking would be a meaningful application of that knowledge.

[Prompt] Would the concept of inert knowledge predict a student might be able to provide a correct meaning for a word from the test if a teacher asked for a definition, but not when the student encountered the word while reading at home?

[AI] Yes, that is a possible example of inert knowledge. If a student is able to provide a correct definition of a word on a test or quiz, but is unable to recognize or use the word in context when reading at home, it suggests that the student has not fully internalized the meaning of the word and is not able to apply it in new or unfamiliar situations. This could be due to a lack of practice using the word in different contexts or because the student has only learned the definition in isolation without connecting it to other knowledge or experiences.

[Comment] The AI tutor is versatile and you can get it to interact with you in different ways. In the content that follows, I ask the service to generate question prompts and then an outline of ideas from the chapter as might be included in a PowerPoint presentation. I then interact with the tutor based on these prompts and parts of the PowerPoint outline. 

[Prompt] Write three discussion prompts based on this chapter.

[AI] Sure, here are three discussion prompts based on the chapter: 

1. What are some examples of boundaries that exist in your own experiences, and how have you crossed them?

2. How might crossing boundaries between different grade levels or schools impact your learning and personal growth?

3. In what ways can using primary sources enhance your understanding of a topic compared to using a textbook?

[Prompt] Write a response to each of these discussion prompts.

[AI] 1. Personally, I found the ideas presented in this chapter to be very relatable to my own experiences with learning. I think that the concept of crossing boundaries between different experiences is particularly relevant, as it highlights the importance of being open to new perspectives and ideas. In my own learning, have found that being willing to explore new topics and engage with different viewpoints has been key to my growth and development.

2. In terms of encouraging students to engage in meaningful learning, I think that it is important for teachers to create a supportive and inclusive classroom environment. This can involve providing opportunities for students to collaborate and share their ideas, as well as offering personalized feedback and guidance. Additionally, I think that it is important for teachers to be flexible and adaptable in their teaching methods, in order to meet the diverse needs and interests of their students.

3. One example of how the thinking processes discussed in this chapter have helped me in my own learning is through my experiences with critical thinking. By learning to question assumptions and evaluate evidence, I have been able to develop a more nuanced and informed perspective on a variety of topics. Additionally, I have found that being able to think critically has helped me to communicate my ideas more effectively, both in writing and in conversation.

[Prompt] What is wrong with reading a textbook from a tablet or computer

[AI] Research has shown that reading from a screen can result in overconfidence and poorer comprehension compared to reading from a paper source. While there are multiple hypotheses as to why this difference exists, some evidence points to metacognition. It is presently unclear why this happens, but external performance and more practice reading long form content from a screen may help compensate for the lack of insight into comprehension. However, it is important to note that this research is not suggesting that reading from a tablet or computer is inherently bad, but rather that it may require more effort and attention to ensure comprehension and avoid overconfidence.

[Prompt] I know that I should try to read from my phone while I am driving. Is this an example of working memory overload that may result in an accident?

[AI] Reading from a device while driving is not recommended as it can lead to distractions and accidents. Research has shown that reading from a screen can result in overconfidence and poorer comprehension compared to reading from a paper source. It is presently unclear why this happens, but external performance and more practice reading long form content from a screen may help compensate for the lack of insight into comprehension. Metacognition, or thinking about thinking, is important for effective tool use and the skilled learner can plan to avoid difficulties or compensate for problems. However, it is important to prioritize safety while driving and avoid any distractions, including reading from a device.

[Prompt] Create the text that would appear on PowerPoint slides used to explain the memory stores and processes and cognitive process that make up the model of cognition used in this chapter.

[AI] Slide 1: Introduction to the Cognitive Perspective

– Describes learning and thinking activities in terms of multiple memory stores

– Processes or mental actions used in thinking and learning

– Executive mechanisms used to evaluate and control processes

Slide 2: Memory Stores

– Three memory stores: sensory memory, working memory, and long-term memory

– Sensory memory holds information briefly 

– Working memory maintains information for a short period of time 

-Long-term memory stores information for a longer period of time

Slide 3: Mental Tools

– Learners have mental tools to accomplish thinking and learning tasks

– Four mental tools: Attend, Link, Elaborate, and Retrieve 

– Attend maintains ideas in working memory

– Link establishes connections between information units – Elaborate adds meaning to information

– Retrieve accesses information from long-term memory

Slide 4: Evaluation and Control

– Learners use executive mechanisms to evaluate and control processes 

– Determine whether processes have accomplished intended goals

– Prioritize certain learning activities based on cognitive constructs

[Prompt] When the slide says the mental tool of attend keeps information in short term memory, I think this means that when we think about something we keep what we are thinking about in our consciousness longer. Is this correct?

[AI] Yes, that’s correct. The mental tool of attend is used to maintain certain ideas in consciousness, which is also known as working memory. This helps to keep the information in our consciousness for a longer period of time, allowing us to think about it more deeply and process it more effectively.

[Prompt] When the slide says sensory memory holds information briefly, just how much time is that? I think it a couple of seconds at most. Is that correct?

[AI] Yes, that’s correct. Sensory memory holds information for a very brief period of time, usually just a few seconds at most. After that, the information either fades away or is transferred to short-term memory.

Summary comment: I hope these examples offer insight into the flexibility of ChatPDF as tutor. Learning how the AI services work takes some experimentation and most services offer the opportunity to experiment without cost. One issue I must acknowledge is that much of what I write about assumes a motivated learner. Thinking is work and work requires interest, resolve, and incentives. None of us is probably motivated to learn some of what others expect us to know. Accepting this reality, I tend to like what Sal Kahn said in defending mastery approaches to learning. Approaches that improve the success of learning address at least an important aspect of the motivational component of education. We are more interested in learning when our efforts result in success than when these efforts are frustrating and often unsuccessful. [my interpretation]

Loading

Chasing a role for computer literacy and coding in K12

One of the issues associated with the advances in the application of digital technology is that the rest of society can’t keep up. For example, advances allow new services and opportunities before the political process can implement laws that redistrict damaging applications. Once applications are widely used, it is difficult to implement constraints that may prove burdensome or expensive for citizens and companies that have invested in new capabilities. 

Education (K12 and higher education) faces similar and related issues. Rapid advances move the workplace faster than educational institutions can generate skilled practitioners and educated consumers to contribute and benefit from new opportunities. In K12, there is a time delay caused by the need to generate expectations (e.g., standards) that guide what exactly should be taught in specific areas of the curriculum and at specific grade levels, and the delay in hiring or developing skills and knowledge in the appropriate faculty members. For long-term commitments, the delay is exacerbated by the time span required to first develop future teachers with the appropriate skill sets and experience.

We closely follow K12 expectations in North Dakota because we worked for many years before retiring in North Dakota and were involved in working with preservice and inservice educators. What is happening in North Dakota may be of little interest to educators in other states. I raise what is happening in ND because education leaders within the state have made the claim that they are the first to mandate, at the insistence of the legislature, universal coding and cyber security experiences for all students.  Even more recent expectations were just established by the state legislature providing another example of politicians requiring commitments from educators. 

One of the components of the ND expectations that may be of use to educators in other states are some of the concrete descriptions of what educational experiences are required of educators. Aside from the standards (which were actually established in 2019, but now have a political mandate for demonstrated application), there are guidelines for educator credentials/certificates. As you might expect, various organizations (colleges and other) are seeing opportunities when only professional development is required and practicing educators need to catch up.

North Dakota CS and Cybersecurity Credentials (guidelines extracted from the document I link to above)

The NDDPI issues three levels of Computer Science (CS) and Cybersecurity Credentials. Each requires a valid teaching license issued by the ESPB.

  • Level 1 – Allows the recipient to teach any CS or Cyber-related course, including advanced courses, at any grade level corresponding with the recipient’s teaching license. Applicants must complete one of the following:
    • 200 or more hours of CS or Cyber training; or
    • 15 or more credits in CS or Cyber; or
    • 3 stacks of micro-credentials in CS or Cyber.
       
  • Level 2 – Allows the recipient to teach introductory and intermediate-level CS or Cyber-related courses, as determined by the department, at any grade level corresponding with the recipient’s teaching license. Applicants must complete one of the following:
    • 40 or more hours of CS or Cyber training; or
    • 6 or more credits in CS or Cyber; or
    • 1 stack of micro-credentials in CS or Cyber.
       
  • Level 3 – Allows the recipient to teach integrated CS or Cyber-related instruction in other contents at any grade level corresponding with the recipient’s teaching license. The recipient also may teach CS or Cyber-related courses in grades kindergarten through grade eight as applicable to the recipient’s teaching license.  Applicants must complete one of the following:
    • 15 or more hours of CS or Cyber training; or
    • 3 micro-credentials in CS or Cyber.

I would argue that while these expectations are being promoted as ground breaking and innovative, the skills and knowledge expected of educators are already lagging behind significant developments in what technology now is and the impact it is having. For example, AI tools now available to all have already changed what coding is and who can code. With the development of ChatGPT, a large language model trained by OpenAI, the relevance of teaching coding in the classroom has shifted. There is still a need for coding education, albeit with a shift towards teaching transferable skills and working with pre-existing machine learning models. Natural language coding is also becoming more prevalent, allowing programmers to write code in a way that is similar to how they would communicate with another person.

For additional reading on the expectations of teaching coding in K12, I would point those interested to this paper in Technology Review. This article provides a history of the emphasis on coding in K12 offering insights into what was taught, who benefited, and how the perceived importance of coding in the classroom has varied over time. Unlike the importance of technology in general that has grown linearly, the interest in coding waned after a peak in the 1980-90s which at the time emphasized vocational opportunities and now has been on an upward trajectory promoting benefits for all again with an expanded emphasis on computational thinking. The Technology Review article suggests that a technology emphasis was often justified as an equity issue encouraging women and minority students to learn the skills necessary to compete in lucrative areas of the economy, but educational efforts were typically disappointing in increasing more equitable access to more financially beneficial employment. 

The applications of technology in society represent an aspect of education that presents serious challenges to the education establishment. 

Loading

Citizen Science Month and Bird Buddy

It is Citizen Science Month. For science educators there are opportunities to recognize the role of citizens in your community contributing to science and perhaps to start a project that allows students to contribute to science.

The Citizen Science website offers explanations and examples, but I have several of my own (personal weather station). Recently, I have exploring the potential of Bird Buddy. My wife became interested in this product as a Kick Starter project and invested. Bird Buddy is essentially a bird feeder that contains a motion activated camera. The feeder uses wifi to connect to your phone.

Images that are shared with you are also sent to Bird Buddy and the company uses AI to attempt to identify the birds in the images that you capture.

In addition, Bird Buddy has initiated a project called HeartBeat that collects information about the birds photographed in different areas of the world (the device is now being sold world wide). As a participant you can access these data and view some of the photos being captured world wide.

An interesting feature of BirdBuddy is that you share images you capture with others you designate. To my knowledge educators have not taken advantage of this opportunity, but it would be possible to collect data and share images with other classrooms.

Loading

Complexity of the textbook cost issue

Capitalism is not a perfect system. No method of encouraging productivity and fair compensation is. Rather than making the effort to sort through the complexities involved in the creation and manufacturing of a product or the delivery of a service, it is too easy to seize on a single issue and feel justified in some conclusion you have reached. I want to argue that this is the case in the public perception of the cost of textbooks.

Why am I writing about this topic? My motive comes from comments made in the reaction to the legal decision made against the Internet Archives and by advocates for Open Educational Resources. My intent to broaden the discussion a bit to address parts of the complexity that are often ignored.

I am the author of a couple of college textbooks and I was a college faculty member. This combination places me in a position of being able to observe both the actions of publishers, but also students, bookstores, and those who comment on various aspects of the book business and the use of commercial instructional materials. A little about me as a textbook author. While I wrote for a small, specific market, I would argue I was successful. The textbook my wife and I wrote had a run of five editions with major textbook publishers. Our book was imagined as a book that undergraduates in education programs would use in a course with a title something like “Technology for teachers”, but was also used in graduate courses and by individuals interested in the topic. We now have the rights to our book because we were unable to work out an arrangement that would offer a $29 Primer in combination with related web content. The idea was that we would update the web contentment continuously editions and organize this content to expand the Primer. In fairness and full disclosure, we wanted to be paid for the continual process of writing rather than a furious revision effort when a new edition was authorized. We now sell the Primer as a $9 Kindle trying to offer a version of our vision.

The issue of textbook cost and what the cost provides:

I have written about aspects of the textbook issue for years. One of my favorite posts was headlined “The beer money ploy”. While the title may seem unrelated, the post explored an aspect of the perceived cost of textbooks that is often ignored. I see parents and politicians talking about the cost of textbooks all of the time. Kids are paying $600+ a semester for textbooks in some fields of study. This seems possible. Our textbook was once sold for between $100 and $140 depending on the supplemental materials that were bundled and books of this type cost nothing compared to books in math and the sciences. Here is the thing about this cost as explained in the “Beer Money Ploy”. This is not the actual cost to the student. Nearly all book stores and many online outlets purchase used textbooks for 50% of the sales price. So a $600 bill at the beginning of the semester allows a resale of $300 to the campus bookstore or online outlet at the end of the semester. If you don’t explain this to your parents, you have $300 to spend as you wish. It is true that bookstores keep an eye on what books have been ordered for next semester and use this when purchasing used books, but the proliferating online services don’t worry about this for a given institution. Used book may be resold several times allowing bookstores and online services easy money for putting books back on the shelf or on the online market.

So, an author or the company paying the author a percentage of the wholesale price to the bookstores have to make their money on the original sale only. This means the company must jack up the price compared to what they would charge for a consumable item. Textbook companies have begin participating in the used market and now may lease textbooks. This solves their problem to some extent.

Textbook companies are not without blame. Textbook companies spend some of the money from sales on selling. Unlike a bookstore, textbook companies pay sales reps that visit campuses and individual instructors. Of course, this contributes to the cost of textbooks but seems necessary to get instructors to take a look at the books they consider. The effort instructors devote to the exploration of the multiple options they have for assignment is a related issue. Bookstores want instructors to continue with existing assigned books. Textbook reps promote their most recent offerings in a given space (large courses almost always encourage multiple books from a single publisher because of the amount of money involved) arguing for the value of current information. While this is true and there be other good reasons for considering a different book, new book adoptions are also how the company, author, and sales reps make their money. See above description of used book market. See previous comment on capitalism.

Publishers

What you pay for with a typical commercial in comparison to a roll your own approach:

A commercial publisher spends money on people who perform functions that may be diminished or absent in self publishing or the absent in the online material an educator might identify and patch together as an information source. When you develop a commercial textbook you work with an editor who comments on everything from writing style, the importance of content included or maybe not included, to embellishments such as when an example, feature box, or chart might be helpful. Sometimes they work with you to cut down the amount of material to meet some cost to page number target. They liked that I put content on the web to support my book because that became an option they could recommend for material I had spent hours to create and they wanted to drop.

Commercial publishers have specialists who check every reference to see that references you include are actually accurately cited at the end of the book and people who specialize in creating an index. They pay photographers to provide images appropriate to the content and people who transform charts and graphs from sketches authors might provide. Pages are carefully laid out translating the page after page of generic text into something that has a professional appearance. Is this worth it? I have different feelings about different things. I am constantly annoyed when I cannot find a reference that an author in a self-published book forgets to add. An appendix of key terms is helpful. The layout I like, but I read a lot of stuff in ebooks that lacks this feature. I read theses and dissertations that are hundreds of pages long and I clearly would rather read something in a more pleasing format. However, I can still read this material and I would not want students to spend time creating content that has a more diverse appearance.

I have already mentioned sales reps who bring free copies of textbooks (another expense) to the office of instructors maybe after they look up what you teach and quickly reading the promotional materials about books they have available on that topic. Necessary? Not in an ideal world, but I also know that few instructors spend time looking through even a few books to select the one they will assign. Some do, many just make a selection based on what they found acceptable in a previous edition, the scholarly reputation of the author(s), or a quick examination of a few topics of personal interest. I admit I tended to switch back and forth between a couple of books I liked so I kind of fall in the middle of this laziness continuum. Being forced to read the book I used at least every other year was my motivation.

One caveat to what I have said. My opinions probably best apply to lower division service courses and less to upper division courses for majors or graduate students. This has to do with the background of the instructor (see following comments), the uniqueness and depth of the content, and what type of literature best suits the purpose of a course. For example, advanced courses are more likely to require exposure to multiple authors who have specific expertise and primary rather than secondary source material.

Authors

Expertise

I would argue that writing a quality lower level, survey textbook in many ways requires more preparation than a specialized upper division textbook. I have found that the survey course requires that I address topics I know should be included that I am not prepared to address. I may be able to write on topics generically because I should know more about most topics than students because of previous teaching experiences, but quality instruction and more so writing instructional content requires a depth of knowledge beyond what ends up conveyed in what is actually written. When you write in an actual area of expertise, the range of topics is much more restricted and you are likely an active researcher/scholar in that area. When I try to explain this to people, I use the example of copyright specifically when it comes to fair use and what classroom teachers can do to present content themselves or that their students have created online. I started reading about this topic and encountered something called the TEACH act (Technology, Education, and Copyright Harmonization). I learned that the purpose of this legislation was to place online and in person instruction on an equal footing when it came to fair use for instruction. Sounds important for K12 teachers to me. To qualify for this equal footing, there were certain expectations – protected access for students in actual classes. By my understanding of what this means, unrestricted access as would be allowed when using popular tools such as web environments (Google Sites) should not be used to share content allowed for classroom instruction because access is not limited, whether intended or not, to students with access. I started asking lawyers with educational responsibility (my university attorney) and experts talking about educational fair use at conferences and they all were baffled by the question and even the existence of the TEACH Act. I have not found reference to this act in other technology textbooks in the discussion of fair use.

This is just an example, but I offer it to make the point that there is some unique work required when creating the background for writing about topics beyond what I would describe as the typical expertise of most authors writing in a broad academic area. There may be far more work required than you realize if you have not written a textbook yourself. Part of what authors are paid for is to do this work. Again, I have a specific kind of product in mind and I assume that support for the statements made and the actions recommended have a basis not in personal opinion but a careful review of expert opinions and research.

Summary

Understanding what is a fair price for work and the cost of a product depends on a realistic understanding of what it takes to generate that product. These comments were intended to communicate some of the factors consumers may not recognize.

Loading

AI in Social Bookmarking Tools

I presently make use of two subscription social bookmarking services (Diigo and Memex) and one service that is presently free (Glasp). If I was forced to select one for its capabilities, it would be Glasp. While I am annoyed when a service I have used for years adds new features I never use and then expects a bump in the subscription price, I have the opposite opinion of Diigo. It does what it has always done, there just have been no new developments in years.

To be fair, most of the features I find most interesting about Memex and Glasp have little to do with the social components. I have written about both services in previous posts so I will not review most features, but concentrate on how Memex and Glasp use AI.

When it comes to AI, my primary interest is how AI can be used to process specific content I have already identified. At present, these two systems differ in how they stored content. Glasp assumes I want to ask questions or ask for the processing of content I have accumulated over time (all bookmarked sources). Memex assumes I want to ask questions of a specific source. There are cases in which each is what I want.

Memex Garden (the computer-based and not the tablet/phone Memex Go) allows the user to apply AI to a selected source (Ask and then select summarize selection, summarize the entire source, or general question). In general web pages are short enough that my highlights and annotations are easy enough to scan that these options are not particularly useful. However, Memex allows the annotation of PDFs which are long enough, and because I mostly read journal articles as PDFs complex enough that summarization may be helpful.

In contrast, Glasp applies AI across all of the content that has been collected. Note, this applies only to web content and not Kindle book highlights and annotations which can be accessed through this service. Glasp does not presently allow the highlighting and annotation of PDFs. The nice thing about the AI application in Glasp is that it identifies the specific sources that are used in responding to inquiries.

While I said I would not focus on social capabilities, Glasp does have an interesting AI social capability. Glasp allows a user to designate the account of another user and ask questions of that user’s public annotations. This may seem a little creepy, but the mission of Glasp clearly emphasizes the collective intelligence or efforts of users.

The following is an example of the content from one of the developers I was able to ask AI to query.

Almost there

I have been able to exchange texts with both the Memex and Glasp developers expressing my interests. I am most interest in having an efficient want to explore the highlights and annotations I have applied to Kindle books and PDFs because these are the sources I mostly focus on in my professional work. I can understand why copyright issues may be factor to consider in sharing highlights from commercial products (books, journal articles), but I would hope there would be an easy way for access to be limited to the sources I own. Perhaps notes could be treated as shareable, but not highlights. Goodreads allows the sharing of some highlights so some sharing must be allowed.

Loading