James Bedford

YouTube video by james bedford AI in Education: Navigating Risks and Embracing Opportunities (Mar 2025)

Full length keynote I delivered on Navigating AI: Potentials and Pitfalls at UNSWC (March 2025). Enjoy.

www.youtube.com/watch?v=uYYn...

AI WHEEL OF DEATH (STUDENT VERSION) based on our 2023 AI Wheel of Death.

If you ask Gemini Deep Research or even ChatGPT deep research you can get a deep overview/report of a person based on their online social media accounts, linkedin profiles, twitter etc. Journalists often use it to get a brief of a person they are interviewing. Careful what you say online :😅

Agreed. I advocate for AI use when used effectively. AI wheel of life to follow. This infograph is to demonstrate what happens when the AI does everything.

Yes. They make it easy to opt out.

Gemini now uses your Google search history to provide specified results beyond just your regular Gemini conversations. Asked it "what do you know about me?" and it knew where I lived, what my interests were and what I'd searched recently in Google.

A circular infographic titled "The AI Wheel of Death" shows a repeating cycle in academic publishing involving heavy reliance on AI. The cycle includes six stages: "Author uses AI to do research" (orange section) "Author gets AI to write paper" (red section) "Reviewer reviews the work using AI" (purple section) "Author responds to reviewers’ AI feedback by asking AI to adjust paper" (blue section) "Journal publishes the paper" (teal section) "LLM scrapes internet during training (which includes this paper)" — leading back to the beginning In the center is a black skull, emphasizing the critical tone of the loop. A smaller skull is at the bottom of the image. The graphic critiques the recursive and ethically questionable overuse of AI in scholarly research and publishing.

The last thing I want is for us all to get stuck in an endless generative AI loop where AI does everything. This is not what I mean when I talk about using generative AI for research and writing. This is lowest common denominator stuff

"AI Wheel of Death" by James Bedford (2025)

Me too but we ran out of tokens. 😂

Copy of GAIL V2.11 Generative Artificial Intelligence Literacy Module (GAIL): Example of an ‘AI-Guided Learning Module’ Prompt Note to Users: The following is the prompt we used for our Generative AI Literacy Module a...

Want to create an AI-guided learning module? Not sure what an AI-guided learning module is? Here's the prompt for our Generative AI Literacy Module (GAIL) released under CC BY-NC 4.0. You can adapt and make your own custom GPT from off these instructions.

docs.google.com/document/d/1...

😎😎😎

What are you doing to prevent skill atrophy because of GenAI usage? Have you noticed your writing, creative problem solving & synthesis skills becoming bit blunt after repeated use of LLMs?

Protecting Human Cognition in the Age of AI
arxiv.org/html/2502.12...

Deep Seek CEO hires people with a humanities background to improve their tech. Anthropic are doing the same. The humanities can help lead this technology into the future and help ensure it is used for good.

www.msn.com/en-in/techno...

#LTHEchat 322: Postplagiarism: Teaching and Learning in an Age of Artificial Intelligence Led by Prof. Sarah Elaine Eaton @saraheaton.bsky.social In higher education, we find ourselves at a fascinating inflection point. The emergence of powerful artificial intelligence tools has forever…

2 days to go before this week's #LTHEchat!
This week we will feature Prof. Sarah Eaton @saraheaton.bsky.social who will be leading a discussion on postplagiarism: teaching and learning in an age of Artificial Intelligence.
Check out the blog for more: lthechat.com/2025/03/10/l...

Nice little AI book stack

Delivered a keynote today for UNSW college sharing thoughts around AI in education. Recently, I find myself talking a lot about the dangers of over-reliance, the decline of critical thinking, and how we might work to reduce the negative impacts that come with using AI tools regularly.

What a 30-Day Break From AI Taught Me About My Teaching Using AI had become second nature for this educator. A month without the tools gave him an opportunity to pause, reflect, and recalibrate.

After a 30-day break from AI tools, educator @jamesbedford.bsky.social feels more dedicated than ever to human-centered learning and the “messy, creative core of what it means to be an educator.” 🧑‍🏫

#EduSky #EduSkyAI

What a 30-Day Break From AI Taught Me About My Teaching Using AI had become second nature for this educator. A month without the tools gave him an opportunity to pause, reflect, and recalibrate.

Really great post by @jamesbedford.bsky.social on @edutopia.org today about his monthlong break from AI tools.

(featuring some neat hero art by the amazing guillemcasasus.com !)

www.edutopia.org/article/down...

What a 30-Day Break From AI Taught Me About My Teaching Using AI had become second nature for this educator. A month without the tools gave him an opportunity to pause, reflect, and recalibrate.

My recent article in @edutopia.org

www.edutopia.org/article/down...

Large language models and data protection IntroductionThe emergence of large language models (LLMs) in late 2022 has changed people’s understanding of, and interaction with, artificial intelligence (AI).

A great explainer article for those wondering about why data privacy matters when using LLMs.

privacyinternational.org/explainer/53...

Dear educators,

Your detailed course outline with 4-page assessment item instructions that students don’t read anyway make for very thorough prompts.

It’s the best. An old classic

a man with a beard is wearing a turtleneck sweater and a scarf . ALT: a man with a beard is wearing a turtleneck sweater and a scarf .

OpenAI might sue DeepSeek for copyright infringement.

www.ft.com/content/a0df...

Introducing Operator A research preview of an agent that can use its own browser to perform tasks for you. Available to Pro users in the U.S.

Soon an ‘AI student agent’ could complete an asynchronous online course autonomously. This isn’t just about using ChatGPT for essays; it’s about handling all course activities from logging into an LMS, and submitting assignments—all w.o any learner intervention

openai.com/index/introd...

Mind you, it took 15 seconds to ‘think’ about it.

If you haven’t tried already, go
and try www.deepseek.com, a new LLM that is said to be on par with o1. It is free and open source. Select the ‘deep think’ option to see the chain of thought the model produces. The quality of outputs I’m getting are some of the best I’ve seen an LLM produce.

I would agree with you but you can access most of the top tier models for free if you know where to look. Deepseek also just released an open source model on par with OpenAI’s o1. I don’t know how it’s possible to make this billion dollar technology even more accessible than it currently is.

Home The AI Pedagogy Project is a curated collection of resources for educators curious about how AI affects their students and their syllabi

Use the tools available. See what they are capable of. Have an open mind. Be critical and rigorous and assume nothing. Do what we’ve always done. But also don’t. Adapt.

Practically speaking this resource is quite good:
aipedagogy.org

When we talk about equity of access and GenAI let’s take a moment to consider this: students can freely access the world’s most powerful LLMs in their pockets (with/w.o internet) 24/7. The equity issue here is providing the support and education to use these tools responsibly.

Unsolicited writing advice, no. #14151:
You don't need to write for publication, or for money, or for an audience, or for anyone but yourself. Joy is a worthwhile objective. Joy doesn't need a reason.

A free text book titled ‘Foundations of LLMs’ has been pre-released via arxiv. It seems accessible for beginners and covers basics to more advanced concepts.

LINK TO NOTEBOOK LM: notebooklm.google.com/notebook/0e513…

LINK TO TEXT:
arxiv.org/pdf/2501.09223