Jeff Greene

For whatever reason, engagement with my posts on @bsky.app has dropped a lot - unless the post is about #GenAI. I get more consistent engagement with my posts on LinkedIn. I'm not blaming this on anything or anyone in particular (besides myself and my posting ability), I just think it's interesting.

Yup. I suspect the bias-adjusted ones are closer to reality.

Learner-Generated and Instructor-Provided Graphic Organizers as Aids to Learning from Text: A Meta-Analysis - Educational Psychology Review Graphic organizers are spatial representations of key textual information—such as matrices, flowcharts, or hierarchies—that can support learning through two primary approaches: as a generative…

Graphic organizers (GO) support memory and comprehension, whether they are learner-generated or instructor-provided. Some publication bias but effects are still practically significant. Efficacy depends on how well they prompt generative learning. #PsychSciSky #AcademicSky #EduSky

"Our overarching recommendation is to avoid reducing responsible educational AI to a process of optimizing technical benchmarks...we advocate for a human-in-the-loop approach to applying LS insights and empirical data to improve learning outcomes...it is critical to involve learning scientists."

Thanks!!

A Smartphone Before Age 12 Could Carry Health Risks, Study Says

I appreciate the article’s emphasis on both (1) this isn’t causal evidence and (2) there’s more than enough associational evidence to support a parent saying, “Sorry, no phone until you are X years old.” One quibble: why can’t I find the actual article online? Where’s the link?
(Gift link)

Kudos!!!

Outrage about scholars using #GenAI to review for them illustrates issues at the individual, social, & structural levels of knowledge creation (i.e., epistemic cognition and systems). For example, at the individual level, offloading the critical role of expert judgment to GenAI violates ethics…(1/n)

Thoughtful essay from a professor at Ohio State wondering why in the world universities are promoting tools of cognitive automation.

Agreed. This is the way, if one insists on inserting GenAI into the curriculum: “The most responsible way for colleges to prepare students for the future is to teach AI skills only after building a solid foundation of basic cognitive ability and advanced disciplinary knowledge.”

two men are standing next to each other with the words be a goldfish on the bottom ALT: two men are standing next to each other with the words be a goldfish on the bottom

Settling in to watch the first Wicked movie and I’ve realized I remember nothing from seeing the show years ago. I’m a musical goldfish.

Comparing the wrong things Search might be a better approach for learning about the world, but the right comparisons matter

I’m no GenAI Zoomer, but I have to begrudgingly admit they’ve gotten less inaccurate over time. @mikecaulfield.bsky.social had a good post on his substack breaking down LLMs v Googling for answers. The differences aren’t as stark as they once were. mikecaulfield.substack.com/p/comparing-...

And I'd be thrilled to see your thoughts included in the higher ed/GenAI thinking, @dannagal.bsky.social . Have you been following @olivia.science and @irisvanrooij.bsky.social 's work in this space? It's very good and important to consider.

...and they are held responsible for any and all errors, etc. That makes me think no one but experts should be using GenAI for knowledge building, and even then, I question its utility. Perhaps for editing? Other than that...I'm skeptical it adds value, at the moment.

Learning With GenAI: Shortcuts Versus Cutting Yourself Short When should you use generative AI to learn? Some say you should use generative AI all the time; others say never. Your answer should depend on your goals.

...I tried to write a bit about these issues here, and have been contemplating writing more. So, the ethics of using GenAI at all are serious and important. In addition to that, for me a big issue is that authors take responsibility for disclosing use... www.psychologytoday.com/us/blog/psyc...

...GenAI for knowledge work v. GenAI for seeking answers as one would do during search. As @mikecaulfield.bsky.social has argued, it's difficult for me to see how using GenAI for answers is different than Googling (ethical issues of not compensating knowledge creators not at all withstanding)...

And, for better or worse, I grant experts more leeway to use GenAI, if they wish, as long as they maintain a strong control position (human in the loop) and they divulge their use clearly. And they are responsible for any and all output. Also, I distinguish...

@dannagal.bsky.social Good questions with no easy answers. I try to separate ML, NLP, and other "AI" techniques, implemented from within an LLM or not, from truly generative techniques that involve little to no expertise on the part of the human. The former are ok, to me, but the latter are not.

Thanks! Means a lot coming from you.

…there are more issues at each levels. What I listed above are just examples. My main point is that looking at this solely as an issue of “bad people using GenAI to offload reviewing expertise” is too narrow a scope to solve the problems. (8/8)

…these levels interact with one another (e.g., structure pressure leads to individuals making bad decisions, which in turn trigger societal responses), we would be wise to focus any solutions on fixes that span individual, social, and societal levels. Finally… (7/n)

…another symptom of a larger structural issue regarding how knowledge is produced and the pressures on experts who participate in multiple roles in that epistemic system. When that pressure gets too great, some people make bad decisions. Not an excuse, just an explanation. As issues at… (6/n)

…of experts, and their graduate students. Review quality necessarily goes down when established experts choose not to review (I’m not saying those who do review are ‘lesser’ - I’m just saying that when you remove good reviewers, review quality necessarily drops.). GenAI use is just… (5/n)

…I’ve chaired program committees at major conferences in my field and, frankly, I was stunned at how many prominent scholars and presenters at conferences were not reviewing for that conference. This results in acceptance decisions being forced on a smaller group… (4/n)

…people’s use of GenAI to review is yet another symptom of a knowledge-producing system under too much stress. I’m not going to excuse experts who offload judgment to GenAI, but this is but another example of such behavior, which has been around for a while now… (3/n)

…At the social level, the outrage illustrates differing levels of tolerance for GenAI as an “agent” versus human-in-the-loop paradigms, with the former far more egregious, but the latter still deeply questioned by some. At the structural level… (2/n)

Outrage about scholars using #GenAI to review for them illustrates issues at the individual, social, & structural levels of knowledge creation (i.e., epistemic cognition and systems). For example, at the individual level, offloading the critical role of expert judgment to GenAI violates ethics…(1/n)

“…At the great public universities that do most of the research and teaching in the United States—and the private universities that are less in the news—the work forges ahead”

My not-so-favorite year As 2025 comes to a close, it’s a good time to step back and assess one of the most tumultuous years in the history of American science. The second Trump administration has brought cuts to so many impo...

I agree with Holden Thorp: there is a foundation to build upon:

“The political discourse—and its coverage in mainstream news outlets—has focused on a handful of mostly elite institutions; a general reader could easily conclude that American science is at a standstill. But…”

Fighting for Truth in a Rage-Driven Algorithmic Age with Jessica Yellin Podcast Episode · On with Kara Swisher · 11/27/2025 · 51m

Interesting @karaswisher.bsky.social podcast about alternative news channels, how channels could curate their partners for more reach, how to stand up your own channel, and the limitations of #GenAI when you care about accuracy and depth.