Educators Technology

Educators Technology Ph.D. in Educational Studies, EdTech blogger, author, founder of ETML & Selected Reads. .

Practical tools and tips about using technology in education, for users, teachers, leaders and managers of educational ICT.

I am reading this old paper with generative AI in mind!Back in 2011, Betsy Sparrow, Jenny Liu, and Daniel Wegner studied...
09/28/2026

I am reading this old paper with generative AI in mind!

Back in 2011, Betsy Sparrow, Jenny Liu, and Daniel Wegner studied what happens to memory when people expect information to remain available on a computer.

In one experiment, participants remembered fewer details when they believed the information they typed would be saved. In another, they were better at remembering where information was stored than recalling the information itself.

I keep thinking about how this question has changed with generative AI. We no longer need to remember where to find an answer. We can ask a chatbot to retrieve information, explain it, summarise it, and put it into words for us.

What happens to learning when the tool can do all of those things in one conversation?

The 2011 study cannot answer that question. Its experiments involved trivia statements and saved folders, long before today’s chatbots.

But it gives me a useful starting point: What do we choose to remember when we expect a tool to be there whenever we need it?

For students, I would ask a related question after an AI-assisted task: What can you now explain, question, or use without reopening the chat?

Reference

Sparrow, B., Liu, J., & Wegner, D. M. (2011). Google effects on memory: Cognitive consequences of having information at our fingertips. Science, 333 (6043), 776–778.

We have been talking a lot about AI and cognitive offloading from the students’ side. How about we look at the teacher’s...
09/28/2026

We have been talking a lot about AI and cognitive offloading from the students’ side. How about we look at the teacher’s side?

What do you lose as a teacher when you delegate lesson planning, feedback, or assessment decisions to AI?

The answer depends on what you delegate and what you still do yourself. Asking AI for a few activity ideas may give you more options to consider.

Accepting an AI-generated lesson without working through why it fits your students is a different matter. If you repeatedly hand over that reasoning, you may have fewer opportunities to develop it, especially early in your career.

In a new review, Chun Sing Maxwell Ho and Junjun Chen examine teachers’ AI-related cognitive offloading. They propose looking at two things: the depth of delegation and the quality of teacher oversight.

A teacher might ask AI to draft feedback, then check it against the student’s work, revise its claims, and adjust its tone. Or the teacher might send it after a quick glance. The same tool and task can involve very different amounts of professional judgement.

The authors are careful about the evidence. Their review found that we still have little direct observation of what teachers actually delegate and how they check AI’s work.

So the question of long-term skill loss remains open.

But it is a question worth asking now. Teaching expertise grows through the work of noticing what students need, deciding what to do next, and learning from the result.

If AI saves us time, I want to know which parts of that work the saved time comes from.

Link in the first comment!

Ho, C. S. M., & Chen, J. (2026). Teachers’ AI-related cognitive offloading: Systematic narrative review and framework for delegation and oversight. Computers in Human Behavior Reports, 24, 101329.

Bloom’s taxonomy has been with us for decades. It has helped teachers think about the kinds of thinking we ask students ...
09/28/2026

Bloom’s taxonomy has been with us for decades. It has helped teachers think about the kinds of thinking we ask students to do, from recalling information to evaluating ideas and creating something new.

In 2001, Lorin Anderson and David Krathwohl led a revision. Among its changes, **Evaluate moved just below Create**.

Now our information environment has changed again. Generative AI has become what I think of as a *semiotic participant*: a non-human system that can produce language and images, respond to our ideas, and keep working with us as we revise them. That kind of participation is now available to almost anyone with access to a tool.

A student can generate an essay, an argument, a lesson plan, or an image in seconds. So what does it mean to place Create at the top when producing something polished has become so easy?

I think we may need to revisit the revised taxonomy. AI can speed up creation, but its output can contain invented facts, weak reasoning, bias, and material with unclear intellectual property status. The finished product tells us less about the thinking behind it than it once did.

Evaluation now has to run through the whole process. Before creating: What is the purpose, and what should remain my work? During creation: Which suggestions should I question or reject? Afterwards: Is the result accurate, fair, and worth sharing?

Perhaps the most demanding task we can give students is to create something and defend the decisions they made along the way.

What would Bloom’s taxonomy look like if we designed it for a world where AI can create with us?

There is a lot of discrimination against scholars from non-English-speaking backgrounds in academic publishing. For year...
09/28/2026

There is a lot of discrimination against scholars from non-English-speaking backgrounds in academic publishing.

For years, many were told to have their manuscripts edited by a native English speaker before submitting them. Now AI has made that kind of language support more accessible, but the judgement has shifted: if their English sounds too polished, it may be treated as suspicious.

Avi Staiman describes this bind in a recent article. Write with noticeable language differences, and your credibility may be questioned. Use AI to improve the prose, and your authorship may be questioned instead.

I also appreciated his point about what can happen during editing. AI may turn “the findings suggest” into “the findings demonstrate.” The sentence sounds stronger, but the claim has changed.

Researchers need to be able to improve the language while keeping control of what their work actually says.

Perhaps journals should spend less time guessing how a manuscript was written from its style and more time examining the research itself. What claims does it make? What evidence supports them? Has language editing preserved the author’s meaning?

Scholars should not have to choose between being judged for their English and being judged for improving it.

Link in first comment!

Have you ever started writing about an idea you thought you understood, then realized halfway through that you didn’t?Th...
09/28/2026

Have you ever started writing about an idea you thought you understood, then realized halfway through that you didn’t?

That moment is part of learning.

In her 1977 essay Writing as a Mode of Learning, Janet Emig argued that writing gives us something unusual: we can see our thoughts on the page, return to them, notice what is missing, and revise what we mean.

Writing also asks us to make connections explicit. A sentence that sounded clear in our heads may turn out to need another sentence, an example, or a different argument altogether.

I keep thinking about this when we discuss AI writing in education. If a student asks AI to produce the paragraph before they have tried to formulate the idea themselves, what happens to the thinking they might have done while writing it?

AI may help a student respond to a draft or find a clearer way to express an idea. But we need to pay attention to when it enters the process. Sometimes the unfinished sentence is where the learning begins.

Nearly fifty years later, Emig's essay still gives us a useful question for designing writing tasks: Which parts of writing do students need to do in order to learn what we are teaching?

Link in the first comment!

Emig, J. (1977). Writing as a mode of learning. College Composition and Communication, 28(2), 122–128.

This is one of the resources I use with pre-service teachers to help them think about curriculum integration.When I ask ...
09/27/2026

This is one of the resources I use with pre-service teachers to help them think about curriculum integration.

When I ask them to connect subjects in a lesson, the first idea is often to choose a theme, say “the ocean”, and find a reading, a science activity, and an art project that all fit.

That can be a starting point, but I want them to ask a harder question: What are students learning from the connection?

Robin Fogarty’s Ten Ways to Integrate Curriculum helps us explore that question.

It shows that integration does not always mean combining several subjects into one big project.

A teacher might make connections within a subject, coordinate related units across two subjects, or carry a thinking skill such as using evidence through several disciplines.

With my pre-service teachers, I ask them to look at a curriculum outcome first. Then they consider which kind of connection would help students understand that outcome more deeply.

If they cannot explain what the connection adds, we go back to the lesson design.

I still find this 1991 resource useful because it gives us a language for discussing those choices. A theme may make activities look connected. The learning goal tells us whether the connection has a purpose.

Link in the first comment!

Reference
Fogarty, R. (1991). Ten ways to integrate curriculum. Educational Leadership, 49(2), 61–65.

We often assume that students who are good at studying will also be good at using AI. But being able to use it and wanti...
09/27/2026

We often assume that students who are good at studying will also be good at using AI. But being able to use it and wanting it involved in your learning are two different things.

In a survey of 484 students at one Australian university, Jason Zagami found that students who reported higher grades were less enthusiastic about generative AI and less likely to say it improved their learning.

Many still used it. They described asking it to clarify ideas, organise material, or help them get started, while checking its answers and keeping the harder decisions for themselves.

The paper calls this "guarded adoption.” I like the term because it captures a position we do not always make room for in conversations about AI literacy: a student can know how to use AI and still decide that a particular task is worth doing without it.

Some students worried that handing over too much would cost them the chance to work through a problem and experience the satisfaction of understanding it. Others wanted their work to remain an honest reflection of what they could do.

For me, this raises a question worth asking in our classrooms: Which parts of learning do students want help with, and which parts do they need to do themselves to feel that the learning is theirs?

Link in the first comment!

Reference
Zagami, J. (2026). Guarded adoption of generative AI in higher education: High-achieving students, successful-student identity, and epistemic agency in a single-university mixed-methods survey. International Journal of Educational Technology in Higher Education, 23, Article 47.

The way we write says a lot about who we are. Our authorial voice, our style, and the way we put our experiences into wo...
09/27/2026

The way we write says a lot about who we are. Our authorial voice, our style, and the way we put our experiences into words all carry something of our identity.

When AI does the writing for us, what happens to that voice?

A new paper by Maryam Khosronejad and colleagues explores this question through three examples of writing with ChatGPT.

The researchers found that AI could produce fluent text while quietly shaping its meaning: favouring some ideas, leaving out other perspectives, and assuming an audience the writer had never specified.

What interested me most is that the writer’s voice appeared in the response to the AI output. The researchers examined what the text was saying, noticed where it missed their intentions, and prompted it in new directions.

Those decisions (i.e., what to keep, question, add, or reject) were part of the writing.

This gives us a useful question for students: Where can we see your thinking in the choices you made with AI?

The paper offers a conceptual analysis of three illustrative examples, so it does not tell us how all students write with AI. It does remind us that a polished paragraph alone may reveal very little about the decisions behind it.

Link in the first comment

Reference

Khosronejad, M., Ryan, M., & Mills, K. A. (2026). Authorial voice in human–machine writing with GenAI: Plays and counterplays of meaning. The Australian Educational Researcher, 53, Article 90. https://doi.org/10.1007/s13384-026-01018-4

Do we know how students are actually using AI, or are we filling in the gaps ourselves?A new study in Frontiers in Educa...
09/27/2026

Do we know how students are actually using AI, or are we filling in the gaps ourselves?

A new study in Frontiers in Education surveyed staff and students in one school at a UK university.

Students were generally more positive about the benefits of generative AI. Staff, meanwhile, had limited visibility into what students did with it outside class.

The student uses that came through most clearly were improving language and producing text.

That matters. If educators mainly see the finished assignment, they may have little sense of whether AI helped a student understand an idea, polish their wording, or produce work the student could not explain.

This was a small survey in one institutional setting, so it does not tell us how all university students use AI. It does give us a reason to ask students about their practices before designing policies around our assumptions.

Link in the first comment!

Kahn, P., Carrigan, M., Wyman, I., Liu, R., Smith, P., Andonegui, A. R., Murtagh, L., & Song, F. (2026). Still emerging: Understanding generative AI use in higher education. Frontiers in Education, 11, Article 1885253.

"You cannot sound like GPT."That's the rule one multilingual computer scientist now follows when editing papers before s...
09/26/2026

"You cannot sound like GPT."

That's the rule one multilingual computer scientist now follows when editing papers before submitting them to a major AI conference.

Haley Lepp and Daniel Scott Smith studied this in their paper "You Cannot Sound Like GPT": Signs of language discrimination and resistance in computer science publishing.

They looked at nearly 80,000 peer reviews from ICLR, one of the most influential conferences in the field, from 2018 to 2024. They also interviewed 14 conference participants from five continents.

What they found is uncomfortable. Papers with more authors from countries where English is less widely spoken got lower ratings, fewer compliments on clarity, and more criticism of their writing.

That held even after the researchers accounted for what reviewers said about the substance of the work.

Many people expected ChatGPT to close that gap. Fix the grammar and the bias goes away. The authors found that the pattern barely moved after ChatGPT arrived.

The interviews help explain why. Reviewers found new clues. Before ChatGPT, missing plurals, tense slips and long sentences suggested a "non-native" author.

After ChatGPT, flowery wording and words people associate with AI did the same job. Polished prose in a "GPT style" led reviewers to the same assumption by a different road.

One interviewee described how her trust in multilingual authors dropped when their writing had errors, and dropped again when it sounded like AI. She laughed as she realized it: "It's damned if you do! Damned if you don't!" (p. 8).

Link in the first comment!

Reference

Lepp, H., & Smith, D. S. (2025). "You cannot sound like GPT": Signs of language discrimination and resistance in computer science publishing. In Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency (FAccT '25). ACM.

Address

Halifax, NS

Website

, , ,

Alerts

Be the first to know and let us send you an email when Educators Technology posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Contact The School

Send a message to Educators Technology:

Shortcuts

Share

Category