cgruenloh

@cgruenloh@hci.social

Senior researcher @ Roessingh Research and Development | Guest lecturer @ University of Twente | Enschede (The Netherlands) | she/her

Passionate about human-centered and value-sensitive #eHealth design that empowers patients, improves quality of life, quality of care and the work environment for health professionals.

#HCI #PatientEmpowerment #ValueSensitiveDesign #HumanCenteredAI #ExplainableAI #SharedDecisionMaking #AIEthics #ResponsibleAI #HCXAI

#Vegan #LongCovid #MECFS #Kindness

This profile is from a federated server and may be incomplete. Browse more on the original instance.

cgruenloh, to academia

Root problem: Demand in research funding is much bigger than supply. Introducing: complicated grant application system to distribute funding.

20xx: Grant proposals are getting very complicated; writing bureaus are increasingly used to support researchers putting their ideas on paper in a way that increases their chances in winning the funding lottery. 1/x

cgruenloh,

2023: ChatGPT is getting used to write grants.

2025: Many many more research grant applications are submitted. So many that this is not feasible any more. Introducing: using AI to filter through all the submitted proposals.

2/x

cgruenloh,

This is inspired by a conversation I had today with someone working at a University grant office. This is not a joke. I learned that for a national grant it’s estimated that 30% are written using AI. And that funding bodies currently discuss whether to use AI to deal with possible exponentially increased submissions?? Besides the “to disclose” or “not to disclose” debate… 3/x

cgruenloh,

SERIOUSLY: What are we doing? 😱
“Solving” a problem (too many submissions) with the very same technology that caused it? And probably making it even worse, because what type of proposals are likely to get through? The human-written or the AI-written ones?

We should look closely at the process; not throwing AI at everything. Only silver lining: closer interaction between applicant & funding body is also discussed. Not sure if that is AFTER the AI filter tho… 4/x

cgruenloh,

In the context of writing and reviewing a letter of reference, this cycle of “writing by AI” - “reviewing b AI” was exactly what @pluralistic predicted:
https://locusmag.com/2023/09/commentary-by-cory-doctorow-plausible-sentence-generators/

… and which is what I thought immediately about when I read this nature article:
https://www.nature.com/articles/d41586-023-03238-5

What a world we live in 😱

5/5

cgruenloh, to random

Great article by @Daojoan that is spot-on on metrics:

“First, metrics fixation narrows people’s identity and self-worth to what they produce. But we are so much more than our output. Wisdom, creativity, relationships and character matter enormously but evade quantification by their very nature.” 1/2

https://joanwestenberg.medium.com/how-productivity-gurus-are-making-you-less-productive-ce271b7cecb6

cgruenloh,

@xahteiwi fabulous article Florian, thx for sharing.

Goodhart's law: When a measure becomes a target, it ceases to be a good measure.

Quote “Why is that so? It’s because once you make the measure a target that has an influence on people (for example, meeting it gets them a bonus, failing at it gets them a demotion), you have wired them to improve the measure, and not necessarily to improve the underlying conditions that the measure originally arose from.”

Spot-on 💯

cgruenloh,

Truly enjoyed reading your post @xahteiwi

The last two steps on the Yankelovich ladder made me think of something related to that worries me for some time:
“The third step is to presume that what can’t be measured easily really isn’t important. This is blindness.

The fourth step is to say that what can’t be easily measured really doesn’t exist. This is suicide.”

Because what can’t be measured won’t end up in AI models and hence not in predictions or decision support.

cgruenloh,

@xahteiwi Argh, I’m sorry 🥲 Totally relatable btw.

cgruenloh,

@xahteiwi Oh wow, that was an interesting read as well; thanks. I always assumed that this question is a proxy for the overall satisfaction. As you said the premise of this question is biased, but its use as predicting growth and then making the score a goal… jeez!

cgruenloh, to random

Great post by @pluralistic

All of this is just a distraction from real and important scientific questions about how (and whether) to make automation tools that steer clear of Granny Weatherwax's sin of "treating people like things.”

Looking studies testing LLM’s against real physicians in terms of empathy and accuracy, I feel that the human part is missing and people are treated as things indeed.

https://pluralistic.net/2023/11/27/10-types-of-people/#taking-up-a-lot-of-space

tante, to random German
@tante@tldr.nettime.org avatar

Wenn du wissen willst, wie doof NFTs sind: Christian Lindner findet sie ein gutes Beispiel für Digitalisierung. (Via @hagen )

cgruenloh,

@tante @hagen ich dachte diesen ganzen NFT Mist hätten wir echt hinter uns gelassen…. 🤦🏻‍♀️🤦🏻‍♀️

Fischblog, to random German
@Fischblog@chaos.social avatar

Ich soll für Spektrum einen Kommentar über nen aktuell veröffentlichten Klimabericht schreiben. Jetzt frag ich mich, ob ich versuchen soll, da noch was Optimistisches rauszudestillieren oder schreib ich einfach was ich denk?

cgruenloh,

@Fischblog was du denkst!!

cgruenloh, to random

Everyone thinks they know how & should work, but very few have actually done .

@mmasnick, Randy Lubin, and Leigh Beadon developed this really fun game, exploring the difficult choices and tradeoffs involved in managing a trust and safety team.

I was immediately hooked and ended with these scores:

🛡️

🏆 Won via IPO and retired from the tech world

📈 Score: 1984
📏 Rating: ⭐⭐⭐⭐◾️

Can you beat me? https://trustandsafety.fun/

danmcquillan, to random

'ChatGPT use shows that the grant-application system is broken' https://www.nature.com/articles/d41586-023-03238-5 - I'd go further; that LLMs can fool bureaucracies with bullshit shows the role of bureaucracy has never been to properly order and regulate, but simply to obfuscate the location of real power

cgruenloh,

@danmcquillan @guihn288

Reminds me of a great post by @pluralistic including the example of professors outsourcing letter-of-reference writing to ChatGPT.

“The next step is obvious: as letters of reference proliferate, people who receive these letters will ask a chatbot to summarize them in a few bullet points, … 1/2

cgruenloh,

@danmcquillan @kdnyhan
Great question! Why do the processes require pages and pages of pointless text? I was asking myself that once i realised that more and more often writing agencies are involved in writing large grant proposal, because the calls are so complicated and you need to know the game….

I think there is something very wrong if researchers are not able to write down their ideas because the process is so complicated and time consuming.

cgruenloh,

@kdnyhan @krisnelson @danmcquillan Yes, sounds about right. Similar to the examples by @pluralistic (professors writing a letter of reference; a lawyer writes a threatening letter), this gatekeeper (or signal that it sends if someone indeed goes through the process) is losing its purpose if it’s done by ChatGPT.

https://locusmag.com/2023/09/commentary-by-cory-doctorow-plausible-sentence-generators/

codingconduct, to random
@codingconduct@hci.social avatar

Starting a human-AI interaction reading group. Anyone has good syllabi, starting points, their own experience to share? 🙏

cgruenloh,

@codingconduct interested in that as well! 📚😉

tante, to random
@tante@tldr.nettime.org avatar

""COVAX and World Bank to Accelerate Vaccine Access for Developing Countries," trumpets a World Bank press release. "How AI Is Making Healthcare More Affordable And Accessible," announces Forbes magazine. "How technology is helping improve financial inclusion around the world," reports CNBC. It's a linguistic frame that appears regularly in media, PR, and policymaking. Those who can't afford the top-tier forms of basic necessities like housing or physical and mental healthcare, we're told, can have "access" to less expensive, lower-quality versions. Enter bottom-rung ACA marketplace plans, less effective COVID vaccines, homeless people living in train containers, scammy cryptocurrency apps, and clunky chatbot "therapists." After all, they're better than the alternative: having no healthcare, housing, or income at all. But why must having nothing at all be the only alternative? Why isn't it possible to ensure high-quality essentials for everyone? And how does media's repackaging of substandard necessities as "increasing access" and fostering "inclusion" serve to make the barbarism of austerity politics seem palatable, even benevolent?"

Very timely podcast episode
https://pca.st/episode/cfe80be6-2cd8-44a3-8e46-f9d27960d044

cgruenloh,

@tante I saw this line of argument when people criticized the launch of an LLM „virtual doctor“ chatbot to answer health related questions (!) which is called Dr Gupta (!!) and launched by non other than Martin Shkreli.
Arguments went from „he did his time in jail, so leave him alone“ to „check your privilege; this makes healthcare accessible“

The latter really made my blood boil. This is not healthcare; this is crap. Which we give to those who we don’t offer high quality care

Fischblog, to random German
@Fischblog@chaos.social avatar

Oh Gott, kann das sein, dass die Schenkelbürste wieder in Mode kommt? 🫣

cgruenloh,

@Fischblog Schneuzer sind auch schon wieder da….
War die Rückkehr der Pornobalken Level 1? Und Schenkelbürste ist Level 2? 😱

twostraws, to random
@twostraws@mastodon.social avatar

Dear men: If you are invited to speak at a conference, it is a huge privilege and I hope you’re excited. But please, please pause before you accept to check the list of other speakers – too many Swift events are happening with the most extraordinary gender imbalance, and it’s bad for everyone. I know you want to travel, present, meet people, and have fun, but please use your privilege to help ensure everyone is included. When we’re building apps for everyone, everyone needs to be in the room 🌈

cgruenloh,

@twostraws went to X only to read the replies. Jeez!! How defensive they are although you only ask men to “pause” and to “help ensure everyone is included”.

Didn’t see anywhere in your tweet that you demanded first time speakers to give up their opportunity. Nice reminder why it’s good that I left that place.

jbigham, to random
@jbigham@hci.social avatar

who wants to put together the Top 100 HCI … we construct it to be as amazing as possible, but then just publish it and then run away. if we do this right, we can fund CHI for the next 40 years off the click revenue.

cgruenloh,

@jbigham @andresmh 😅 😅 😅

cgruenloh, to random

Are you at Mensch & Computer and interested in in design? Michael Bui will present our paper on “Patient Journey Value Mapping: Illustrating values and experiences along the patient journey to support eHealth design” today at 14:00 in building 4.

https://dl.acm.org/doi/10.1145/3603555.3603558

cgruenloh, to ai

is not problem yet? / are just amazing tools? All good; except beware of AI taking over on the future?

@garymarcus illustrates with several examples in this post that the problem of LLM is already very much visible:

„The Imminent Enshittification of the Internet: LLMs are creating a huge sanitation problem that will probably never be solved“

https://open.substack.com/pub/garymarcus/p/the-imminent-enshittification-of?utm_campaign=post&utm_medium=web

cgruenloh, to ai

Großartiger Podcast: Marc Raschke im Gespräch mit @tante

Wozu entwickeln wir eigentlich Technologie? Um es den Menschen leichter zu machen und es ihnen besser geht oder um ihre Lebens- und Arbeitsbedingungen zu verschlechtern?

Offenbar bin ich auch ein Luddit; denn für mich sind Menschen, ihre Bedürfnisse, Gesellschaft, Werte und Umwelt zentral beim Design von Technologien.

https://zuendholz.podigee.io/8-new-episode

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • JUstTest
  • mdbf
  • everett
  • osvaldo12
  • magazineikmin
  • thenastyranch
  • rosin
  • normalnudes
  • Youngstown
  • Durango
  • slotface
  • ngwrru68w68
  • kavyap
  • DreamBathrooms
  • tester
  • InstantRegret
  • ethstaker
  • GTA5RPClips
  • tacticalgear
  • Leos
  • anitta
  • modclub
  • khanakhh
  • cubers
  • cisconetworking
  • megavids
  • provamag3
  • lostlight
  • All magazines
    •