Lightnews — Scholar-powered news

Anna Mills

@annamillsoer.bsky.social

6.3K followers 3.4K following 1.2K posts

Writer, community college writing teacher, obsessed with AI in education, #OER advocate, author of HowArgumentsWork.org. annarmills.com

Posts Media Videos Starter Packs

Pinned

Anna Mills @annamillsoer.bsky.social · Nov 10

Hey, here's my starter pack of folks I follow on Bsky around AI, in relation to writing, tech, regulation, thinking, and pedagogy. It's a mix of critical, enthusiastic, and moderate voices. Who did I miss? go.bsky.app/NG7WrA5

56 37 150

Anna Mills @annamillsoer.bsky.social · 1d

Tomorrow!

From Passive to Active: Teaching Students to Critically Engage with AI Feedback

A free webinar through AI for Education

Tuesday, October 7 at 12 pm ET
www.aiforeducation.io/from-passive...

From Passive to Active: Teaching Students to Critically Engage with AI Feedback — AI for Education

Many educators worry that AI feedback will make students passive consumers of suggestions rather than active thinkers and writers. But when students learn to critically engage with AI feedback—questio...

www.aiforeducation.io

1 3

Anna Mills @annamillsoer.bsky.social · 1d

Thank you, I hope it's helpful!

Reposted by Anna Mills

Margaret Mitchell @mmitchell.bsky.social · Aug 6

🤖 ICYMI: Yesterday, @hf.co and OpenAI partnered to bring open source GPT to the public. This is a Big Deal in "AI world". Allow me to explain why. 🧵
huggingface.co/openai/gpt-o...

2 17 54

Anna Mills @annamillsoer.bsky.social · 11d

Building Your Own Multi-Pronged Approach to Reducing AI Misuse in Writing Assignments

For the California Virtual Campus Online Network of Educators

Slides: link.annarmills.com/building
Recording: onlinenetworkofeducators.org/webinars/
Worksheet: link.annarmills.com/worksheet

I'd love feedback...

Webinars - Online Network of Educators

onlinenetworkofeducators.org

1 1 4

Anna Mills @annamillsoer.bsky.social · 16d

Beyond Citation: Describing AI Use in Your Work from the MLA Style Center @mlanews.bsky.social sends the message that "Indicating where and how AI has touched your research and writing is just as much a part of academic integrity as citing your research sources is."
style.mla.org/describing-a...

Beyond Citation: Describing AI Use in Your Work | MLA Style Center

Learn about acknowledging substantive use of AI in your work

style.mla.org

2 9

Anna Mills @annamillsoer.bsky.social · 18d

Why we need guardrails/accountability measures/assessment reform.

www.theatlantic.com/technology/a...

I’m a High Schooler. AI Is Demolishing My Education.

The end of critical thinking in the classroom

www.theatlantic.com

2 4

Anna Mills @annamillsoer.bsky.social · 18d

Is anyone posting on here about AI agents accessing learning management systems and completing work for students without students even knowing what the assignment was?

Big problem.

Here David Wiley shows ChatGPT Agent in Canvas www.linkedin.com/posts/openco...

#EduSky #AIineducation

"ChatGPT completes online homework in 3 minutes: A demo" | David Wiley posted on the topic | LinkedIn

People are talking about ChatGPT's Agent Mode and how it might be used to complete online homework, but I haven't seen anyone share a specific prompt or a video of it working. In the demo below, I use...

www.linkedin.com

1 2

Anna Mills @annamillsoer.bsky.social · 22d

Building Your Own Multi-Pronged Approach to Reducing AI Misuse in Writing Assignments

This Thursday, September 18, 2025; 1:00-2:00 PM Pacific Time

Free and open to all through the California Virtual Campus Online Education Initiative.

Register: onlinenetworkofeducators.org/webinars/

Webinars - Online Network of Educators

onlinenetworkofeducators.org

1 1

Anna Mills @annamillsoer.bsky.social · 22d

So moved that someone took the time to nominate me for an Open Education Global (OEGlobal) Award for Excellence! How wonderful to be on the list with colleagues from so many countries who are part of this global movement for open education.

awards.oeglobal.org/2025-nominat...

2025 All The Nominations

As reported at the close of this year's nomination cycle, the countries represented in this year's nominations are: Angola, Argentina, Australia, Austria, Bahamas, Brazil, Burundi, Canada, Cape Ver...

awards.oeglobal.org

Anna Mills @annamillsoer.bsky.social · 24d

Love it! Just got a LInkedIn message addressed to "Hi {{FirstName}},"

Thank you, that's even easier to ignore!

Anna Mills @annamillsoer.bsky.social · 24d

That's true... he did clearly love the humanities and wanted to be clear-eyed and honest with us about what's happening, which I appreciate.

Anna Mills @annamillsoer.bsky.social · 25d

In the interview, I got the sense that he just wasn't interested in trying to preserve anything; he wanted to be one of the ones thinking about all the cool new things that are possible.

1 1

Anna Mills @annamillsoer.bsky.social · 25d

The part about making the humanities more about how to live resonated for me too.

But it seems like just gives up on educational experiences that exclude AI, it seems. That doesn't make sense to me.

1 1

Anna Mills @annamillsoer.bsky.social · 25d

I fully agree about the loneliness of it. But there's also an element of boredom I haven't been able to articulate the reason for. I do think AI books could be generated to challenge us in way we want to be challenged. But that's not satisfying if there's no human intention or vision behind it.

Anna Mills @annamillsoer.bsky.social · 25d

Not on board. AI-generated books are not equivalent to human-written books.

Human intention matters to readers.

"[T]he making of books such as those on my shelves, each the labor of years or decades, is quickly becoming a matter of well-designed prompts..."

Will the Humanities Survive Artificial Intelligence?

Maybe not as we’ve known them. But, in the ruins of the old curriculum, something vital is stirring.

www.newyorker.com

2 3

Anna Mills @annamillsoer.bsky.social · 25d

Thank you! Intriguing

1 2

Anna Mills @annamillsoer.bsky.social · 25d

Playing with @jonippolito.net's app What Uses More to compare generating an AI image with streaming Netflix for an hour or storing 5 GB in the cloud.

How can we get away from unnecessary streaming and cloud storage, since they use a lot more than most AI except video generation?
what-uses-more.com

What Uses More? Compare the environmental footprint of digital tasks

Compare the energy and water footprint of AI and digital activities.

what-uses-more.com

1 3 8

Reposted by Anna Mills

Dr. Damien P. Williams Wants A Better Sociotechnical Paradigm @wolvendamien.bsky.social · 29d

How many times, in how many contexts, from how many internal and external researchers, or from how many CEO's are people going to have to receive this message before they believe it:

"Hallucinations" are an inherent part of the large language model architecture.
www.nature.com/articles/d41...

Can researchers stop AI making up citations?

OpenAI’s GPT-5 hallucinates less than previous models do, but cutting hallucination completely might prove impossible.

www.nature.com

18 200 660

Anna Mills @annamillsoer.bsky.social · 29d

Thanks! I just know we are all wondering when we read anything these days...

Anna Mills @annamillsoer.bsky.social · 29d

Interesting! I didn't find any guidance about that and so assumed I shouldn't, but I did wonder if it would be okay to upload it to a platform that wouldn't use it for training. In the end, I decided I didn't need it anyway in this case; I preferred to use my own traditional processes.

Anna Mills @annamillsoer.bsky.social · 29d

@leoflores.bsky.social I appreciate you're looking at these questions!

Anna Mills @annamillsoer.bsky.social · 29d

When you peer review, do you indicate somehow whether you used genAI in the process?

I'm peer reviewing a book and put this AI Use Statement at the end of my comments: "I used only Grammarly spelling and grammar suggestions and no generative AI in the preparation of this peer review."

3 6

Anna Mills @annamillsoer.bsky.social · 29d

There are definitely reasons to keep searching. Chat-assisted research doesn't have to take over. However, chat assistance can help with information verification. Traditional search can't do the sorting that might help us see what kinds of sources line up to support or debunk a claim.

Anna Mills @annamillsoer.bsky.social · Sep 8

We can do better, Bluesky! Let's stay civil, even if you hate AI and someone is posting about something that can be done with AI.

Mike "looking for research fellowship" Caulfield @mikecaulfield.bsky.social · Sep 7

Why I block people after posting something controversial like "you can get better at using AI if you stop thinking of it as an answer machine"

Anna Mills @annamillsoer.bsky.social · Sep 8

To get a chatbot to do a better job of evaluating a claim, try @mikecaulfield.bsky.social's "sorting prompts" (with web browsing turned on):

"Read the room: what do a variety of experts think about the claim?

OR

What is the evidence for and against the claim I posted?"

Mike "looking for research fellowship" Caulfield @mikecaulfield.bsky.social · Sep 7

This is a good post on how to get better at using AI. It's the product of pretty intensive research on what consistently works although I don't surface that research directly in it.

Bluesky and AI, so obviously comments on this are off. mikecaulfield.substack.com/p/is-the-llm...

Is the LLM response wrong, or have you just failed to iterate it?

Many "errors" in search-assisted LLMs are not errors at all, but the result of an investigation aborted too soon. Here's how to up your LLM-based verification game by going to round two.

mikecaulfield.substack.com

1 3