About this blog

This is the blog of Neil Ingram and reflects a variety of my interests over the years. As a biology teacher, university academic, examiner and author.

I have been deeply interested in the use of IT in schools when Windows 3.0 came out in May 1990.

About ten years ago, I was invited to be part of the Hewlett Packard Catalyst initiative and I developed a model about how school pedagogy could work in the Web 3.0 world. Some of this has been published, but quite a lot has not.

The development of Artificial Intelligence systems is generating the same levels of excitement as Windows 3.0. As the CEO of Microsoft said, “It feels like the 1990s again!“.

I have developed an introduction to the thinking:

Pedagogy AI is a roadmap for the pedagogy of a lesson using AI with students.

This is based on the ideas in a series of ten linked posts, called Teaching and learning with AI. Part one is here.

To show you round the rest of the site:

Stories from Nowhere was a lockdown project, trying to use stories to bring important ideas into middle years biology lessons. It is built on observations in a wood and work we were doing on a 16-19 curriculum framework for the Royal Society of Biology

Exploring the epigenetic landscape is a microsite about how genes interact with the environment and uses the ideas of Conrad (Hall) Waddington.

The home page contains a mixture of posts relating a university course I ran on genetics, society and education.

There are also posts on Evolution relate to a book I co-authored for Oxford University Press.

The title “tools for clear thinking” is based on a book by Conrad (Hal) Waddington, whose ideas run through every article on this site. The cover image was designed by Dall-E3, and the prompt included Waddington’s term “epigenetic landscape. I was delighted that its rendering resembled the original conception by the painter John Piper.

Pedagogy AI

Download a high resolution .png image here. Download a powerpoint slide here.

Pedagogy AI is a roadmap for a lesson in which students use AI for themselves. It is based on observations of lessons where students used iPads to research and solve problems.

This is a versatile pedagogy. There are times when the teacher teaches by direct instruction, with visible, evident, controls (eg stage 01). At other times, the teacher gives the students some level of autonomy (eg Stage 03), and yes more invisible controls. Even so, the teacher is always on the sidelines, ready to facilitate and encourage.

In this pedagogy, students are taught to use AI safely and productively. There are clear lesson aims, specific learning outcomes, which the students discuss together with the appropriate format for the output from AI. During the prompt-writing stage (03) the students work with the evident support of the teachers. They are encouraged to experiment and make mistakes, which are points for reflection and further refinement.

The outputs from AI are shared and evaluated critically (04). The best answers can be combined to contribute to the discussion of the problem. Sometimes the problems have a real world context and a variety of different viewpoints, such as “should we vaccinate young children against diseases like measles”, or “how can we live more sustainably?”. A balanced account of all of the issues and views could be appropriate learning outcomes for such tasks.

As students become used to this way of working, they will gain increasing confidence and autonomy (04). Even so, behind this independence, the teacher remains a supporting critical presence.

Finally (and most importantly), students reflect privately on their learning from the lesson (05), reflecting not only about the problem under study, but also their experiences of using AI.

A word on the symbols:

This represents a teacher-centred pedagogy, such as direct instruction. The teacher is using explicit and visible controls over the content that is being taught and the way that the lesson proceeds.
The teacher is giving some control and autonomy to the class to direct the course of the lesson. The teacher  is using more invisible controls, but is still present as a facilitator at the side of the lesson.
In this stage, the students are working independently, but under the supervision of the teacher. This allows the students some autonomy, whilst ensuring that the students work safely and productively.
This represents students working with increased confidence and autonomy, whilst still under the support and guidance of the teacher. This is an example of a versatile pedagogy. The teacher can intervene with visible controls, when necessary. 

Read about the ideas behind this road map in a series of blog posts on this site, Teaching and learning with AI (parts 1-10).

Go to the first article here.

Teaching and learning with AI (part 1)

An explosion of creativity

an explosion of creativity with AI

Teachers are discovering new ways of using AI in their teaching, in an explosion of creativity.

Their joyful enthusiasm is infectious! Teachers can use the many examples in the table below to inspire or plan lessons for students of all ages and abilities.  

The ability of AI to simplify and translate text into other languages allows teachers to meet the diversity of needs of their students more effectively than ever before. Generative AI is a game changer. 

The examples are loosely grouped into categories, depending on how they impact on lessons. The learning conversation includes all of the interactions in the classroom and where learning takes place. The learning community describes how the wider community (including online support) can influence teaching and learning. The classification is somewhat arbitrary, as some examples could appear in more than one group. 

Using AI to “reduce workload across the education sector [and] free up teachers’ time, allowing them to focus on delivering excellent teaching” (Department of Education, England, October 2023) are aims that are shared by many education policy makers across the world. 

Students could also use most of the examples shown above for themselves to promote their own learning. The official policy documents are rather more tentative about this, generally recognising that, for example: 

“The education sector needs to:

  • prepare students for changing workplaces
  • teach students how to use emerging technologies, such as generative AI, safely and appropriately” (Department of Education, England, October 2023)

One of the biggest concerns for contemporary pedagogy is to consider how students can use AI in the classroom to promote active and productive learning. This is rather more than using AI to plagiarise homework assignments.  

This is not without its risks. Students using AI in the classroom have the power to disrupt the normal functions and flow of the classroom, and this is something we need to consider next.

Q. Which areas of your workload do you think that AI could help you to manage more productively?

Q. What are the advantages and disadvantages of teaching students to use AI productively?

Updated: 11/01/24 to include link to Microsoft’s “Unlock generative AI safely and responsibly” – a new classroom toolkit from Microsoft Education”. A resource for 13-15 year old students.

Updated: 16/01/24 to include end of article questions.

Next article

Teaching and learning with AI (part 2)

AI the disruptor

AI has the potential to disrupt classroom life

AI is increasingly being described as a “key disruptor” of business and commercial systems. In this, it is following a well-established pattern of the effects of any new technology on society. “We shape our tools and thereafter our tools shape us,” as Marshall McLuhan (almost) said.

The phrase “key disruptor” is used less often when talking about education, especially in schools. This is because its iconoclastic power is constrained and regulated by a number of other powerful forces. 

Until now, the use of IT has been absorbed into the existing structures of education: rarely does it reshape the equilibrium of school life, which (for many) is built on routine and regular external examinations and inspections.  

Even so, it is worth remembering that IT and AI in particular, fit Kranzberg’s first ‘law of technology’: ‘technology is neither good nor bad; nor is it neutral’. 

This series of articles look at the powerful forces that constrain the use of ICT and AI in school classrooms. They use the important ideas of visible and invisible controls proposed by the English sociologist Basil Bernstein. 

This series of articles (numbered 1-10) is intended to be a contribution to initial teacher education and the professional development of teachers.

Background: teacher-centred and student-centred pedagogies

Educational theory is still dominated by the split between traditional “teacher-centred” learning and progressive “student-centred” learning. These are often treated as polar opposites and mortal enemies

The twin approaches are shrouded in ideology. From 2010, there was a shift in educational policy in England towards a “knowledge-rich” curriculum. The architects of this change often cite ED Hirsch as a founding father. 

Hirsch promotes a teacher-centred pedagogy of a knowledge-rich curriculum. Unsurprisingly, he is highly critical of John Dewey, whose book “Experience and education” is a high watermark for student-centred learning.

Teacher-centred learning uses teacher’s questions to stimulate memorisation of information, promote understanding and critical thinking, as in a “Socratic dialogue”.  

This is named after the teaching of the greek philosopher Socrates, who taught in the Lyceum is Athens.

Student-centred learning encourages students to build their own understanding, often in social situations and was favoured by the Russian psychologist Vygotsky and the American psychologist Bruner. It is the basis of inquiry-based learning, problem-based learning and project work. 

Traditional and inquiry-based learning are often presented as being discrete categories but can also be seen as being complementary approaches, with different strengths and weaknesses.

Seeing the approaches as opposite ends of a spectrum, allows us to imagine many intermediate positions occupying the centre ground. These blend the best of the two approaches. 

These are dynamic, versatile pedagogies, because teachers can move between pedagogies depending upon how the lesson develops. 

Teacher-centred and student-centred learning become different tools in a teacher’s toolkit, to be used at different times and circumstances. 

The pedagogy chosen will influence how AI will be used in schools. To see how, we need to explore one of Bernstein’s most important ideas: visible and invisible controls. 

Q. Think about a typical lesson in your school. When do you give control to students to:
* work independently
* work in groups?

How do you ensure that the students know what to do and stay focussed on the task?

Previous article                                       Next article

Updated 16/01/24 to include end of article question.

Teaching and learning with AI (part 3)

Visible and invisible controls

Basil Bernstein realised that teachers in traditional classrooms used very visible methods to organise and control their teaching. He called these ‘visible’ controls because they were continuously made explicit to students throughout the lesson. 

These controls affected every aspect of the teaching process including: 

  • the selection and choice of lesson content
  • the use of appropriate subject language
  • the learning outcomes
  • the sequence of presenting information
  • classroom seating and organisation
  • the use of resources
  • the ways in which students listen and answer questions

Traditional classrooms place the responsibility for these controls exclusively in the hands of the teacher, who is the sole voice of authority in the room. 

In student-centred learning, (such as problem-based learning) some, or (very occasionally, all) of these controls are devolved to students, who are given some autonomy in deciding how the lesson will proceed.

These are “invisible” controls because they are implicit and often negotiated between the students and teachers. Group work may be favoured, whereas traditional classrooms firmly insist on individual work.

Notice that the teacher is still responsible for maintaining “invisible” controls. Control is not absent in inquiry-based learning, it just operates in different ways. 

The process is a dynamic one and a teacher may make some of these controls more visible if the students are struggling to achieve outcomes for themselves. Likewise, a strictly visible control might be relaxed to give students increased responsibility for controlling their learning. 

Q: Think about when you use visible and invisible controls in your classroom. How much control do you give your students in these parts of classroom life:

* the selection and choice of lesson content
* the use of appropriate subject language
* the learning outcomes
* the sequence of presenting information
* classroom seating and organisation
* the use of resources
* the ways in which students listen and answer questions

Previous article                                           Next article

Updated 16/01/24 to include end of article question.

Teaching and learning with AI (part 4)

ICT and visible and invisible controls

Laptops, tablets, wifi and modern software are designed to encourage mobility, collaboration, and group work. 

Surfing the Internet is a divergent activity, bringing together into a single page a diversity of ideas and a range of views from across the web. Often these ideas are conflicting and of differing standards of reliability. Reconciling the reliability and value of different web pages is a high-level skill that many students find hard. Thus, the internet is a potential source of confusion and distraction: a disruptive force in classrooms. Even in those classrooms that favour invisible controls. 

“Eyes on teacher”

In its unrestricted use, ICT can disrupt the order of traditional classrooms and potentially undermine the authority of the class teacher. It is not surprising that teachers in such classrooms want to impose controls on its use. 

In 2013, when laptops were issued to all students in some schools in New York, they came with “a one-touch classroom-control feature to lock their screens, replacing whatever was on them with an eye symbol and the phrase “Eyes on Teacher. ”

Similarly, in the earliest versions of Nearpod (a mobile app), the presentation slides displayed on the students’ devices could only be changed by the teacher. These are clear examples of visible controls of the sequence and pace of the lessons. 

(Later versions of Nearpod did allow teachers to programme some levels of student autonomy into their activities. But notice how it is still the teacher setting the controls.)

Even so, the use of ICT in traditional classrooms is often restricted to projections onto a whiteboard, sometimes using a visualiser, allowing the teacher to remain firmly in control. 

Some teachers use managed ecosystems, like Microsoft Teams or Google Classroom, which give students access to the lesson notes and resources when they are working away from the classroom. This relaxes the visible control over where learning can take place.

The use of ICT in student-centred learning is usually more extensive as the teachers teach the students to use software to develop and organise their work. Early stages of this may involve explicit instruction from the teacher, showing that, in practice, inquiry learning is versatile enough to move between visible and invisible controls as the need arises. 

Group work can extend beyond the classroom: students can work together in groups using social apps and this can be facilitated by the school using Teams or Google Classroom.

Q. Think about a successful use of ICT in your lesson. Why do you think it was successful?

Q. Think about an unsuccessful use of ICT in your lesson. Why do you think it was unsuccessful?

What could you do to make it better?

Previous article                                             Next article

Updated 16/01/24 to include end of article questions.

Teaching and learning with AI (part 5)

AI and controls

ICT networks are known to change society in several predictable ways. Rainie and Lee suggest that: 

“in networked societies, boundaries are more permeable, interactions are with diverse others, connections shift between multiple networks, and hierarchies tend to be flatter and more recursive*.”

Rainie, Lee; Wellman, Barry. (2012) Networked: The New Social Operating System (The MIT Press).

In a recursive hierarchy, people have different positions in different teams: a person can lead one team and be a junior member of a second team and an observer of a third team. The hierarchies are flatter because there are fewer levels of authority and there are greater opportunities for collaboration and communication between members of the group.

School classrooms tend to have a flattened hierarchy, with a teacher at the top and the students below being of equal lower status . 

AI is now a member of our network and will impact on the hierarchies, relationships and the boundaries that currently exist. So, how could AI affect and be affected by visible and invisible controls? 

The artificial boundary between visible and invisible controls breaks down when considering AI. 

Imagine the preparation for a lesson where a teacher is using some training prompts to train AI for the students to use. 

The teacher is using the prompts to place, explicit (visible) controls on AI. This is because the prompts are making explicit demands on AI.

Within the lesson, a student using the AI is not aware of these controls, which are acting invisibly.

If these controls give the teachers the confidence to allow their students to use AI in independent study, then the controls are also acting invisibly to the student. 

In these circumstances, the use of AI can become a controlled element in any lesson. The use of training prompts as controls could minimise the potential for distraction and loss of teacher authority, minimising the risk of AI producing errors or hallucinations. 

OpenAI give instructions for writing good prompts with AI. These include giving AI a persona, such as “You are a teacher of English to secondary age students, specialising in Shakespeare’s play “Romeo and Juliet”.

It then suggests breaking the task into smaller steps, similar to how teachers would approach the task with a class.

Give AI a clear indication of the appropriate language and the learning outcomes, as well as how the work is to be presented.

In other words, teachers should prepare AI in the same ways that they prepare their students for the task. AI is not the replacement school teacher in the room, it is just a talented student, acting as a learning support.

AI could be used to promote learning in both traditional and progressive learning classrooms if it was prepared to a sufficient level of compliance to schools’ academic and pastoral requirements

This is not to say that students should not be challenged to critically evaluate the answers given by AI. It is imperative that they do, in the same way that they should be challenged to critically evaluate the responses given by a search query on Google. It is much easier to do this when students have some mastery of the subject being studied, and this may impact how it is used in traditional teacher-centred classes.

Q. “Writing good AI prompts can help teachers think about how they might approach teaching a lesson with a class.”

To what extent do you agree with this statement?

Previous article                                             Next article

Updated 16/01/24 to include end of article question.

Teaching and learning with AI (part 6)

How students can use AI in the classroom 

The recent announcements by OpenAI and Microsoft that organisations can now build their own customised chatbots is something of a game changer for education, because it enables the pre-training that I have discussed in earlier articles in this series to be implemented and monitored. This would enable a version of AI to be produced that students can use effectively in the classroom because it is stable and reliable. 

The remaining articles will look to see how students could use AI in classrooms. We will focus on four key aspects of teaching and learning, which occur in all lessons in all classrooms. 

The four areas are: 

The learning context concerns the selection and use of appropriate knowledge and skills as content for the lesson.

The learning outcomes consider the various ways that students become aware of the criteria for a successful piece of work.

The learning conversation discusses varied interactions in the classroom that lead to students’ learning the lesson content. This includes social conversations and the private conversations that a student has with their “inner voice” that lead to learning.

The learning community considers the impact of the wider community beyond the classroom that impact on teaching and learning.

ICT and AI can affect each of these areas in a variety of different ways, as we shall see in the next articles.

Q. Think about the different kinds of learning conversations that take place in the classroom. How do you help and encourage to have their own “private” conversations, which is where the learning takes place.

Can AI help students with these private conversations or is it solely another form of social conversation?

Previous article                                             Next article

Updated 16/01/24 to include end of article question.

Teaching and learning with AI (part 7)

The learning context

The knowledge and skills that are taught in a lesson have a special context. They are part of a wider sequence of ideas that are intended to lead a student towards mastery of the knowledge or skill. The context is further defined by a variety of local and national factors, as shown below.

The knowledge and skills are appropriate to the age and stage of development of the student, building upon what the student already knows or can do. This is true of classrooms across the world. 

The selection of age-appropriate content has “official” approval by governments (such as the English National Curriculum), or federal state or local district or by the school itself. Professional bodies (such as the Royal Society of Biology or the American Association for the Advancement of Science) also make important contributions to defining what is age-appropriate content, such as the Common Core standards for English and Mathematics in the US. 

It is not surprising that AI-powered apps that are pre-trained on these standards are emerging as tools that teachers can use to plan lessons. Applications such as Teachally will generate lesson objectives, resources, assignments and exit tickets (see this for an example tutorial for a maths lesson). Support for problem-based learning is also available (see tutorial on Treasure Island).

Khanmigo, developed by the regarded Khan Academy, has a lesson planning tool, as well as important modules for student learning that we will preview later.

Both of these applications have a US focus but are likely to develop more international versions in time.

For subjects with external examinations or key tests, the producer of the examinations is also a major stakeholder in defining ‘approved’ knowledge and skills. ‘Approved’ in this case being answers that earn marks in examinations. Such is the pressure to achieve high examination results, that the mark schemes of individual examination questions are often the effective de facto standard for approved knowledge, at least for students aged 14-18. This means that the examination boards (or awarding bodies as they are currently called in England) could play a very important role in developing AI chatbots for the students studying the courses built on their examinations. 

AI pre-trained with the essential policy documents produced by these organisations could be able to answer this student question, which ChatGPT currently finds difficult to address:

Q: what do I need to know before I start learning this new topic? 

A: a self-assessment checklist would be provided with an option to find out more about each of the items. 

The value of this query in allowing students to prepare for new topics before they are taught could be considerable.

The learning context is about allowing students to learn to recognise the language used in the classroom and to be able to use it for themselves. This can be in oral conversation or in writing or in actions. It will soon be possible to have a verbal conversation with an AI chatbot, and this opens up a significant potential for learning, especially new languages. 

Bernstein argued that many students are excluded from learning in schools because they simply cannot recognise the context of the learning: either the words that are being communicated by the teacher or because they cannot form appropriate written or oral responses. 

Teachers of vocabulary sometimes use Isabel Beck’s Three Tier model of vocabulary as a framework for understanding how words in the English language can be categorized based on their complexity and commonality. The AI in education  post ‘Prompt: Mastering Vocabulary with Isabel Beck’s Tiered Model and Comprehensive Teaching Strategies’ shows how teachers can use AI as a vocabulary instructor using Beck’s model.   It a good model for how training prompts can be crafted.  

Teachers often say that a student ‘knows the ideas but cannot get it down in writing’. AI does offer the potential for giving students unlimited writing practice and support for increasingly complex writing.

An appropriately trained chatbot, able to give constructive formative feedback, could be an enormous benefit to students at all ages of their learning.  Khanmigo purports to offer constructive feedback on academic essays and support during the learning process.

Microsoft offers a reading coach with their immersive reader as one of their learning accelerator tools. Microsoft co-pilot tools built into their office applications (such as Word and Powerpoint) may also be useful in helping students to use the language of the classroom.

Q. The learning context defines two important aspects of teaching:
* the appropriateness of the context
* the use of appropriate language and behaviours to include all of the students in the learning.

Assuming you have trained AI to recognise these aspects, how might the “simplify” and “translate” functions of AI help you to prepare for a mixed ability class with learners with individual needs.

Previous article                                             Next article

Updated 16/01/24 to include end of article question.

Teaching and learning with AI (part 8)

The learning outcomes

Paying close attention to the learning context empowers a student to enter the learning and to respond in appropriate ways. This is not just by using technical language correctly, but also by giving responses that are appropriate to the social environment of the classroom and school. 

A student who tries to use AI in an environment where it is explicitly banned or even implicitly discouraged, may feel that they are not entering the learning environment of the classroom. They may feel like an ‘outsider’. 

Recent data suggests that about 1 in 5 teens in the USA who have heard of ChatGPT have used it for schoolwork. If this is widely applicable, then schools do need to find positive ways of supporting AI use by students, even if it does not directly feature in classroom teaching. 

Learning outcomes can be negotiated with students, and students could be encouraged to declare their use of AI by filling in a simple evaluation of what went well and what could be done better. 

Sharing the learning outcomes for a particular piece of work with AI would allow the student to get immediate direct feedback on a draft piece of work, which would then be completed and handed in for teacher marking.

An appendix of help given by AI in the production of the work and its feedback would help to legitimise the private use of AI by students and bring it into approved classroom practice.

In traditional teacher-centred classrooms, AI can be a guest, but this is at the discretion of the teacher as to whether it is welcome in a particular lesson. These are part of the visible controls of the learning outcomes. 

So many of the skills that AI is good at, such as translation, summarising, brainstorming, synthesis, writing accurate prose, composing stories, interpreting scientific data, are skills that the teacher wants their students to acquire on their own. Therefore, there may well be occasions when a lesson should not include elements of AI. 

A counterargument is that, since these tools now exist, why not allow students to develop these skills alongside their use of AI. There are parallels here with the adoption of calculators on the development of mental arithmetic skills. 

In England, this tension has been partially resolved by designating certain parts of the curriculum for the development of mental arithmetic.

Whatever the decisions made during lesson planning, it would be appropriate if these expectations were visible at the outset of the lesson, to avoid the embarrassment of students feeling uncomfortable. 

It would be easier if these expectations were visible at the outset of the lesson, to avoid the embarrassment of students feeling uncomfortable. 

In inquiry-based lessons the teacher acts as a “facilitator.  AI can act as a co-facilitator, with the students evaluating the usefulness of any suggestions made. 

Students could be encouraged to draft and test prompts to get appropriate responses from AI. In a sense this is a digital equivalent of the “think-pair-share” technique, widely used in schools. 

From these discussions, the shape of the task can emerge. However, it is important that the conversation continue until the learning outcomes are clearly established for all students. 

Sometimes a student-centred activity can be surrounded by activities with more visible (teacher-centred) controls. Students using AI to brainstorm important ideas for a project can work autonomously.


This could be more productive if the activity was preceded by a clear discussion of the expected context for the project and its learning outcomes.

Regular opportunities for the student to discuss their thinking with the teacher can also help to keep the student on track. This is an example of a versatile use of visible and invisible controls.

Q. Imagine you are planning a lesson in which students will use AI to research a topic. You have prepared AI by giving the appropriate learning context and learning outcomes.

How will you prepare your students to use AI in the task? What kind of instructions would you give them and how would you monitor the output given to the students?

Previous article                                             Next article

Updated 16/01/24 to include end of article question.

Teaching and learning with AI (part 9)

The learning conversation

Learning takes place within the learning conversation. Broadly, there are two kinds of conversations: social conversations and private conversations. 

All learning has a social context. A teacher asking questions and dealing with answers is a learning conversation. So are the conversations in group work. Even reading a book is a social conversation: readers are responding to the author’s voice.


The second learning conversation occurs inside the learners’ heads. It is their private learning conversations. Memorisation of new information and developing connections with previous learning are part of this learning conversation. This is essential if established learning is to take place.

Teachers plan their lessons around social learning conversations. In traditional teacher-centred classrooms, the selection of resources, and the sequencing and pacing of activities are under the teacher’s visible control. The use of AI in such lessons would be as a lesson episode, a component of a lesson. 

For student-centred lessons, students are given more autonomy over the flow of the lesson and when and how AI would be used. The teacher’s role would be to facilitate this, ensuring that students use AI responsibly with critical awareness. 

AI could have several features that strengthen inquiry-based teaching: 

  • the value of immediate accurate feedback
  • the ability of students to control the pace of their learning
  • the ability of students to address ideas that they find difficult by asking questions. 

AI’s ability to simplify, translate and generate new questions is potentially significant. 

In a Socratic dialogue, questions such as:

  • What do you mean by that?
  • How do you know?
  • Can you give me an example?
  • What are the consequences of that?
  • What is the counterargument? 

Students could be trained to use such questions as the basis for their responses to the suggestions made by AI. 

An excellent example of this is the AI in Education post by Craig Donoghue: AI-Assisted Socratic Learning: Refining A-Level Physics Explanations. Here students are encouraged to engage in a dialogue with AI, asking for questions, giving responses and obtaining feedback. The teacher becomes a facilitator for the learning. The lesson shows “how AI can be a powerful ally in personalized learning”. 

The article gives a sample prompt: 

“I am a year 12 student of A level Physics in the UK. I am revising the topic of X, and am struggling with coherently explaining Phenomenon Y. If I give you the relevant sections from the specification I am studying, can you ask me questions that require explanations only on Phenomenon Y. I will then provide a written explanation myself. I would then like you to evaluate the strengths and weaknesses of my explanations and guide me using a Socratic questioning method towards improving my written responses. Before we begin, ask me any questions you need to fully understand what I’d like us to do.”

One advantage of doing this in a classroom is that teachers can refocus the lesson with some explicit instruction if the group is struggling or extend it further if the lesson is being successful. 

The private learning conversation is where students actively engage with the lesson ideas, asking questions, memorising, questioning and testing their understanding. This often takes place after the lesson, in private or with small groups of friends.

It is here that a trained AI has the greatest potential to be of value. If it were stable enough to act as a “thinking companion” it could be of real value to the metacognitive development of students’ learning.

Examples are already starting to appear. Khanmigo offers AI support alongside its resources. 

Khanmigo is using AI imaginatively, too. It enables students to “chat” with historical or fictional figures (from Elizabeth Bennet to Zeus!), which might be of value for project activities. 

A key insight is that teachers remain responsible for ensuring that AI does not do the thinking for the student, either through lazy plagiarism or through some deeper psychological or emotional dependence.

Controls can be added to AI to prevent students circumventing the thinking by asking for the answer directly. This image from Khanmigo illustrates the type of controls that can be used.

Khanmigo responds by asking a question intended to get the student thinking.

AI can also respond to students’ questions, by: 

  • simplifying ideas
  • organising ideas into mind maps
  •  generating and marking test questions
  •  generating quiz cards as retrieval practice 

to promote the private learning conversation. Sooner or later, however, the student must actively commit ideas to memory. No amount of technology can take that final step.

Q. Do you think students could benefit from having a “thinking companion”?
If this companion were to be AI, what characteristics would you like it to have to be a helpful “thinking companion”?

 Previous article                                             Next article

Updated 16/01/24 to include end of article question.