Meryl Krieger

@Merylk
87 Followers
290 Following
92 Posts
I research, talk about, and teach on the impact of digital technology on culture (especially AI these days), do instructional design, and am passionate about: baking, cats, crocheting, and running, in no particular order. Writes about stuff.
#genAI #generativeAI
I'm down the rabbit hole today, looking at the studies available for accuracy rates on generative AI summaries. For those of you on a similar journey, I came across this https://www.cjr.org/tow_center/we-compared-eight-ai-search-engines-theyre-all-bad-at-citing-news.php
There was the #BBC study in 2024 (reported on in their Feb 2025 article (https://www.bbc.com/news/articles/c0m17d8827ko) but since that was a whole year ago, I'm compiling more recent reports. Please do share in comments any I'm missing? Yes, I have the report from January 2025 about the Apple AI summary
AI Search Has a Citation Problem

We compared eight AI search engines. They’re all bad at citing news.

Columbia Journalism Review

@404mediaco 3/3
This brings to mind all sorts of industry-level research on their own products designed to say that the product might have issues but is "fixable"
- biggest of all, there is absolutely no discussion here in the study on automation bias. I would've expected you all to call this out

I expect better from you all - I depend on your articles to do more than summarize a study. This being said, please keep up the work you do!

@404mediaco 2/3
think -why is AI called that instead of CIP? AI was labeled in response to the dialogue around cybernetics research in the 1950s, and CIP was considered "clunky")
- the study itself is not a great example, and there's a lot of question around the agenda of the sponsoring agency. Note that only one person (of course, the lead author) is not a Microsoft Research employee?
@404mediaco 1/3 This is an interesting but hugely problematic study (yes, I've read it, cover to cover). Let me count the ways (this'll take a couple of comments since I have a lot to say here):
- Bloom's taxonomy is a hugely problematic way of identifying critical thinking skills. It's been massively contested - for decades - particularly since it's getting used out of context

New: We got the pitch dek for the company that wants to target ads by listening to your devices. Google has cut them off in response to our reporting:

https://www.404media.co/heres-the-pitch-deck-for-active-listening-ad-targeting/

Here’s the Pitch Deck for ‘Active Listening’ Ad Targeting

404 Media previously reported Cox Media Group (CMG) was advertising a service that claimed to target ads based on what potential customers said near device microphones. Now, here is the pitch deck CMG sent to prospective companies. Google has kicked CMG off its Partner Program in response.

404 Media
@fullyabstract @Gargron I completely agree. I've been teaching online for years now - in the early days my recordings were *so* stilted, but after much practice it's gotten a lot easier. I keep my audience in my head (I have a good idea how people will react)
This. And, yes. My cats have been the saving of me so many times I've lost count. #CatsOfMastodon https://www.theatlantic.com/newsletters/archive/2024/05/a-man-and-his-cat/678357/
The cat who saved me

I will never owe another cat the debt that I owe her.

The Atlantic
@mwichary super excited to get my copy! Thanks for the newest update.
@duaneaubin I'm familiar with the idea. I'd agree that the commoditization element is something that is particularly US-based, but also that there's less of a governmental filter. European nations are far more hands-on here for a whole host of reasons, whereas in the US, the capitalist mindset is really driven by the Big Tech/Silicon Valley approach. So much to say about this - I'd love to hear what brings this up for you?
@[email protected] it always struck me as a very different storyline and focus after #GodEmperorofDune - you're not alone.