Steve Carney

🚨 New paper alert 🚨 The explosion of consumer-facing AI raise questions about who is more likely to adopt it. We examined how people’s knowledge of AI impact AI receptivity. Against most people’s expectations, it's not tech experts! And not for the reason you may be thinking🧵

Great idea, thanks @mhusseinlab.bsky.social ! Loving these start packs to simplify network building.

Meme with woman making a grossed out face because of a 5/5 star rating with 19 ratings next to the same woman making a pleased face because of a 4.6/5 rating with 2,280 ratings. Text: “I think this meme demonstrated the importance of sample size better and more efficiently than any science class I’ve ever taken”

Reposting for the very satisfying data visualization

Sounds very helpful! Are there any data protections in place / implications for putting other’s IP through this model? Not sure I’m comfortable with my work flowing out to the greater model for training.

A very basic one you’re probably used to from Linux, but just in case: Spotlight search (Command + Space) opens *everything*

Eight misconceptions about nudges. (Sad and surprising that they continue to circulate, even among distinguished scholars.) @katymilkman.bsky.social @geowu.bsky.social
papers.ssrn.com/sol3/papers....

Academic poster presentation providing an overview of the linked manuscript. The key finding is that disclosing content as AI generated or “Made with AI” reduces engagement because it reduces connection to the creator.

Great time sharing my work with @stephtully.bsky.social and Ignacio Riveros at @sjdm-tweets.bsky.social! We find people engage with AI disclosed content less; not because it’s worse quality, but because they feel less connection to the creator. Pre-print here: papers.ssrn.com/sol3/papers....

Saw someone refer to it as “X-Twitter” today and will exclusively be referring to it as such from now on.

@dggoldst.bsky.social has both a Marketing Academics and a JDM starter pack filled with great folks!

My LinkedIn is particularly filled with *new insights* like: “did you know too much choice might be a bad thing?!”

Instead, it’s a protracted exposé vilifying a few trying to do something proactive about the field’s “atmosphere of complacency.” Articles like this are exactly the “risk” he reports on that leads to this atmosphere.

What the article fails to discuss: any of the structural incentives (or lack thereof) that might correct an unquantified problem. Flashy effects get you publications, publications get you tenure. What doesn’t get you tenure? A self-driven internal audit of your data provenance.

The Business-School Scandal That Just Keeps Getting Bigger The rot runs deeper than almost anyone has guessed.

I am curious what the b-school discussion of this article is/will be. I find it thought provoking but deeply problematic — and who is the audience? Probably not me as the article paints a picture of rampant fraud and casts the whole field aside as trivial.
www.theatlantic.com/magazine/arc...

Happy consumer sentiment switch day!

a database of datasets ⬇️

been working for awhile on pulling together economic datasets for econ & policy students b/c there is *so* much publicly available data

sharing it for anyone who finds it useful

pls suggest additions; work in progress

lenorepalladino.notion.site/Economic-and...

Amazing! Thank you!

I had low expectations, but the starter packs have refreshingly invigorated my feed. Particularly @dggoldst.bsky.social’s #jdm and #mktsci packs, but I’m biased 🤷‍♂️

Title, author information, and abstract of working paper.

🚨New Working Paper🚨
w/ Ignacio Riveros and @stephtully.bsky.social

How do AI-generated content disclosures applied to online content impact engagement?

TikTok data and experiments show disclosures decrease engagement, but not b/c ppl find the content worse...(1/2)

papers.ssrn.com/sol3/papers....

(2/2)
Why do AIGC disclosures lower engagement? It’s not because people see AI-generated content as lower quality or inauthentic. Consumers feel less connected to the content creator when AI is involved, decreasing engagement.

Check out our full paper here: papers.ssrn.com/sol3/papers....

1) A date-time picker that can be customized (e.g., MM-DD or HH:mm) and a currency text entry field without the need for Javascript or Regex validation. Bonus points for customizing how that data is recorded.

2) Ability to host files >16mb.

How funny is ChatGPT? A comparison of human- and A.I.-produced jokes: http://osf.io/5yz8n/