Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Dr. Casey Fiesler Profile picture
Dr. Casey Fiesler
@cfiesler
Professor who is on Twitter much less than she used to be. See pinned post for where to find me. information science professor @cuboulder. PhD/JD.
2 subscribers
Mar 30, 2023 • 17 tweets • 4 min read
This is a bad headline. But the findings themselves (about teen girls' perspectives of how multiple social media platforms impact their well-being) are interesting! Let me tell you what they actually are. 🧵 washingtonpost.com/education/2023… This report (commonsensemedia.org/research/teens…) is based on a demographically representative national (U.S.) survey of 1,300 girls age 11-15, in which they were asked about their perceptions of how different social media platforms and specific design features impact their well-being.
Dec 17, 2022 • 11 tweets • 3 min read
In which Santa's elves learn an important less about AI bias. 🎅🤖🧵

'Twas the weeks before Christmas
And Santa was busy
His list was so long
It was making him dizzy

There are too many children
List checking can’t scale
If I do this by hand
I will certainly fail!

1/n But one of his elves
Had a brilliant idea
What if we automate
Christmas this year!

There always are kids
Who are naughty or nice
Every year for millennia
You’ve checked that list twice

Every child there’s been
From Aaron to Zeta
We can build an AI
Because we have the data

2/n
Dec 10, 2022 • 7 tweets • 3 min read
The @huggingface GPT detector works very well on ChatGPT-created text. I ran 5 student essays and 5 ChatGPT essays for the same prompt through it, and it was correct every time with >99.9% confidence. huggingface.co/openai-detecto…

So re: deception use cases (e.g. cheating)... ... it's also against OpenAI's TOS to pass off outputs as being generated by humans. So though people keep telling me "it will get 'smart' enough to fool the detectors!" I would expect @OpenAI to release its own detector in response to ethical concerns.
Dec 9, 2022 • 6 tweets • 3 min read
So I give a lot of advice to PhD applicants. My video about statements of purpose has 60k views on YouTube. Anyway, I asked #ChatGPT to write a statement of purpose for admission to our PhD program. And... it was pretty darn good. But there are some VERY important cautions: The first version I gave it was very specific: our PhD program, a specific type of research experience and interest, and the names of three specific faculty members. It came back with this as part of the statement. If I read this I would 100% think it was a real applicant. But... Professor Fiesler's research on the ethical implications of
Dec 8, 2022 • 4 tweets • 2 min read
ChatGPT says that its own creation was unethical, and then makes a number of suggestions for mitigating harm, none of which appear to be in place. (Maybe the TOS suggestion, which is the least useful one anyway, if cheating falls under the broad umbrella of "deception.") Is it ethical to build an AI that students could use to cheaWhat measures could be taken by the developers of an AI chat Since ChatGPT does say that it is "trained to decline inappropriate requests" and one of its second suggestions for mitigating cheating might have fallen into this category, I fed it 10 algebra problems from a worksheet I found online. It gave great answers to all of them!
Dec 8, 2022 • 6 tweets • 2 min read
On the topic of students using ChatGPT, this is an essay assignment for my tech ethics & policy class. ChatGPT picked "I, Robot." I then used the same prompt but specified "Ex Machina." Decent essays... with identical final three paragraphs, including fabricated citations. So... ASSIGNMENT 2: AI IN SCIFI ...One potential approach to t... If one student thinks of using ChatGPT to write their essay by giving it the assignment prompt, another student might as well. Resulting in nearly identical essays. Also you'd have to find real sources to cite to replace the bogus ones.
Nov 21, 2022 • 8 tweets • 3 min read
Do you think that the ethical reputation of a company impacts whether graduating computer science students are willing to take a job there?

This was the topic of @CUBoulder CS alum Ella Sarder's honors thesis, now published as a poster at #CSCW2022. Here's what we found... 🧵 This was an exploratory study (and we're working on a larger-scale survey as a follow up!); she interviewed 12 graduating students about factors they consider in the job search, how they define a “good” or “bad” company to work for, and how ethics education impacts their choices.
Nov 6, 2022 • 4 tweets • 2 min read
ok so
it appears that the new strategy for advocating for changes to Twitter's content moderation policy is to do things that directly impact Twitter's owner
this seems... not ideal @elonmusk going forward, an... Also I've been thinking a lot lately about how it's harder to imagine harms to people not like yourself. This is a particularly good example of not seeing a potential harm (and thinking about how to mitigate it) until it impacts you.
Nov 6, 2022 • 14 tweets • 4 min read
I'm playtesting #Mastodon explainers. Let me know if this makes sense please. :)

Imagine if your gmail account meant you could ONLY email gmail users. That would be silly! Especially if you don't want Google to have all your data. So why are social networks like that? (1/?) Mastodon isn't a platform- it's a protocol for anyone to run a platform that connects to others. So e.g., I'm on an "instance" set up by a prof friend, but my feed (like your Facebook feed or your Twitter feed) includes content from anyone I follow across any instance. (2/?)
Nov 5, 2022 • 5 tweets • 2 min read
I think this represents a common misconception about Mastodon! The entire point of federated social media is that it isn't a silo. On Twitter you are in a Twitter silo. On a Mastodon instance you can interact with anyone on ANY instance. That said, right now Mastodon *feels* like a silo because there aren't many people there - so yep, it isn't a global community. But neither was Twitter at the beginning. If we can only be on social media where everyone else *already* is, then we'll be held hostage forever.
Nov 4, 2022 • 5 tweets • 3 min read
The ML ethics team at Twitter has done amazing work. e.g. they were *auditing their own systems* and *publishing papers about it* and generally making me feel better about everything happening there.

Anyone should know by now that firing your ethics leads is not a good look. Like seriously, @ruchowdh is one of the best in the biz and Twitter gained a huge amount of credibility with this. blog.twitter.com/en_us/topics/c…
Oct 28, 2022 • 17 tweets • 4 min read
Two years ago we published a paper about online community migration, and I think the findings and recommendations are relevant to what I'm seeing right now on Twitter. Especially the barriers and challenges to relocating (and maybe some solutions). 🧵cmci.colorado.edu/~cafi5706/CSCW… Moving Across Lands: Online... Here are some bad things that are likely to happen when people pack their bags and leave a platform:
(1) content loss (which *might* be less of a big deal on a social platform like Twitter compared to a content creation platform like YouTube)
Aug 28, 2022 • 16 tweets • 4 min read
"The goal of a [program committee] has become to destroy rather than to develop." sigbed.org/2022/08/22/the…

Please bear with me; I'm about to compare the toxic culture of rejection in computer science peer review to a high school orchestra competition. 🧵 A friend told me a story about a low income high school orchestra participating in a competition. A cellist who doesn't own her own cello & had never had a private lesson started a solo. After 1 minute a judge (a college music prof) said "you can stop, I don't need to hear more."
Aug 27, 2022 • 7 tweets • 3 min read
A TikTokker (@ curt.skelton) has convinced a *significant* number of people that he’s actually been AI this entire time and I think this is an interesting window into people’s perceptions of the capabilities of AI. (This video has 12 million views.) 🤖🧵 I was always surprised by how many people thought that the “I forced a bot” things were real, but on the other hand, we’re certainly a farther along in creative AI than we were five years ago.
Jul 16, 2022 • 10 tweets • 3 min read
I would like to share with you the actual narrative of an actual published paper :-\
- LGBT people have privacy concerns so they might be reluctant to participate in research
- So we built a machine learning classifier to identify them on Twitter so they can be researched anyway Ethical considerations:
- The data is public
- Our IRB approved it
- We are not disclosing PII
- We have co-authors who identify as LGBT
Jun 10, 2022 • 5 tweets • 4 min read
Since my stack of books has gotten so much bigger: reading recommendations on tech ethics and justice, the sequel. List for this round:
Data Feminism (@kanarinka @laurenfklein)
Atlas of AI (@katecrawford)
Behind the Screen (@ubiquity75)
System Error (@robreich @mehran_sahami Jeremy Weinstein)
Technology of the Oppressed (@DavidNemer)
Programmed Inequality (@histoftech)
Jun 10, 2022 • 15 tweets • 4 min read
I just saw a comment on TikTok that Planned Parenthood performed 400 million abortions last year. Just presented as absolute straight up data. That is like three times the total number of female humans in the United States. I don’t have a SoundCloud but go follow me on TikTok so I don’t feel so bad about spending enough time on that app to stumble across this kind of thing. tiktok.com/@professorcasey
Apr 13, 2022 • 7 tweets • 2 min read
In my online communities class yesterday students wrote fictional news headlines about online communities and social media 1 year, 5 years, 10 years, and 30 years in the future. Some of my favorites... 2023:
"People take to Reddit to stop World War III"
"Truth Social sued for defrauding investors"
"Editing is out, reality is in: A timeline of the decline in photo editing on Instagram"
"Putin joins Tinder"
Mar 4, 2022 • 15 tweets • 6 min read
#BlackMirrorWritersRoom is a teaching exercise for ethical speculation in computing. @shamikalashawn & I interviewed instructors who have used it, and came away with great insights about benefits, challenges, and cautions in this new #SIGCSE2022 paper! 🧵 cfiesler.medium.com/the-black-mirr… First, it turns out that (even though sometimes they're shocked to be asked to do it!) the opportunity to be creative and think through ethical issues is really well received by students, even in otherwise technical classes. And what they come up with is amazing!
Jan 28, 2022 • 16 tweets • 7 min read
TL;DR: for-profit mental health data sharing from a CRISIS HOTLINE. Read the thread below from @Iwillleavenow or if you prefer you can watch what my capslock frustration actually sounds like. Another excellent thread on this what I imagine will be a popular ethics and privacy case study for the future 😢
Jan 24, 2022 • 4 tweets • 1 min read
Random commenter: Why does CS need an ethics focus?
Me: Because technology hurts people and maybe if doesn’t have to.
Them: Technology is just a tool.
Me: That. That is why CS people need more ethics focused training.

I swear this is 100% real. I’m so tempted to do a video reply to this comment on TikTok because I don’t think everyone realizes that this is an actual attitude. Torn between two trending sounds: “the woman was too stunned to speak” or “this is a joke right?”