What is Reflection?
On technology-assisted exploration and change
I’m Tim Gorichanaz, and this is Ports, a newsletter about design and ethics. You’ll find this week’s article below, followed by Ports of Call, links to things I’ve been reading and pondering this week.
Among the chatbots and code generators in the current wave of new generative AI tools are a bunch of efforts meant to help people with self-reflection.
Me.bot, for example, is meant to help with personal development. The iPhone has, since 2023, included the built-in app Journal, which “helps users reflect on everyday moments,” powered by machine learning. Researchers, too, have for many years been examining how AI can help enable reflection. A paper from researchers at Microsoft presented at this year’s CHI Conference on Human Factors in Computing Systems studies AI-assisted reflection during work meetings.
Then there are the artists. Perhaps what art does best is stimulate reflection. “Latent Reflection,” by Rootkid, presents the notion of a sentient AI trapped within a Raspberry Pi minicomputer, while Dr. Suzanne Kite’s exhibition Dreaming with AI at the Museum of Contemporary Native Arts invites reflection on the material components in AI systems.
And in education over recent decades, reflection has become a byword for good teaching. We want our students to reflect because reflection is good for you. I myself often quote John Dewey: “We do not learn from experience, we learn from reflecting on experience.” So of course some are writing about how generative AI might help students reflect.
But what is reflection? And how do we do it?
All of my classes involve some form of self-reflection assignments along the way, and I’ve noticed that most students do not spontaneously deliver the kind of depth of reflection I’m looking for.
But what do I mean by “depth” of reflection? What even is reflection? Saying “I know it when I see it” might suffice for the U.S. Supreme Court, but it doesn’t do much as a teaching method.
Understanding Reflection
One framework for understanding reflection that I’ve found useful comes from a 2010 human–computer interaction research paper by Rowanne Fleck and Geraldine Fitzpatrick, presented at the Australian CHI satellite conference.
In this paper, Fleck and Fitzpatrick wrangle the vast academic literature to offer some axioms about reflection:
First, reflection always happens with a purpose, so one’s purpose in reflecting will influence the nature of reflection. It follows that if someone has no particular purpose for reflecting (as when a teacher tells you to reflect, or else), then the reflection will be low-quality or nonexistent.
Next, reflection requires time, guidance and encouragement. Again, it won’t just happen spontaneously.
Finally, reflection can vary in depth (the following descriptions taken mostly verbatim from their paper). A person must step through these like a ladder—you can’t skip steps.
R0: Revisiting – Description or statement about events without further elaboration or explanation. Not reflective, but a vital step toward reflection.
R1: Reflective Description: Revisiting with Explanation – Description including justification or reasons for action or interpretation, but in a reportive or descriptive way. No alternate explanations explored, limited analysis and no change of perspective.
R2: Dialogic Reflection: Exploring Relationships – A different level of thinking about. Looking for relationships between pieces of experience or knowledge, evidence of cycles of interpreting and questioning, consideration of different explanations, hypothesis and other points of view.
R3: Transformative Reflection: Fundamental Change – Revisiting an event or knowledge with intent to re-organise and/or do something differently. Asking of fundamental questions and challenging personal assumptions leading to a change in practice or understanding.
R4: Critical Reflection: Wider Implications – Where social and ethical issues are taken into consideration. Generally considering the (much wider) picture.
As a note, reflection may be self-reflection (thinking about one’s self and identity), but it can also be reflection on a specific topic, such as whatever one has been learning in a course or at a conference. Many of the AI tools emerging today focus on self-reflection, though that’s only a small part of the opportunity space.
Facilitating Reflection
Reflection is something all of us humans can naturally do—though again, with the appropriate purpose, time, guidance and encouragement. What I mean is reflection doesn’t require any fancy tools.
Fleck and Fitzpatrick cite Jennifer Moon’s book Reflection in Learning and Professional Development in listing numerous non-technological ways that we can help each other reflect and do so better: writing exercises, prompt questions, dialogue, reviewing past materials, using evaluation, etc.
But of course technology can support reflection, and it stands to reason that the right technologies could augment our natural ability to reflect, making us superhuman as in Steve Jobs’ image of “a bicycle for the mind.”
In their paper, Fleck and Fitzpatrick provide some thoughts on how technology could support reflection, some of which we see in the latest AI efforts:
prompting a person to provide explanations and justifications
bringing in other people to enrich reflection—which is often better in a pair or small group rather than solo
collecting more data than a person could perceive unaided (e.g., video from a different perspective, sensor data presented in an infographic)
Interestingly, Fleck and Fitzpatrick suggest that technology shines in helping people with reflecting on the levels R0–R2—revisiting and exploring relationships.
But they are skeptical that technology could play a strong role in R3 and R4 reflection. That’s because these are about changes and considerations fundamental to each person. “This is not to say that technology will not have a role to play in the actual practice of transformation,” the authors write, “but that arguably the main role for technology is in supporting the foundational resources and processes of reflection.”
Put differently, on their account, if a technology were to attempt to facilitate R3 or R4 reflection, it might produce the appearance of such reflection, but it wouldn’t really be that kind of reflection—because the essential structure of it is that you have to do it yourself.
If you disagree, then there’s your invitation to build something or conduct a study to explore this further.
If you agree, then there’s your reminder that however much you might use or create an app or whatever to help you reflect, the tool can only ever walk with you part of the way.
Ports of Call
Library vinyl club: The Free Library of Philadelphia has a new vinyl club, headed by Drexel LIS alum Jane Lippman. Since 2006, vinyl record sales have been climbing every year; though they’ll (probably?) never reach the heights they did in the 1970s, in 2024 more vinyl records were sold than any year since the late 80s. An interesting example of the return to analog we’re seeing in many parts, even in this digital age.
A winner’s race report: Last week I ran the Western States Endurance Run, coming in some 15 hours after the winner. For a very different race report to mine, check out winner Caleb Olson’s account of the race.