#CriticalAI
LightNews LightNews
burdsandwurds.bsky.social
if ai can’t even get a fast food order right, there’s no way in hell you should trust it with medical advice.
A photo of a McDonald’s drive through with two social media posts on top of it. The first reads “BBC News (UK) @BBCNews McDonald's removes Al drive-throughs after order errors Critical Al read.dukeupress.edu/critical-ai/issue” The second reads “@CriticalAI Let's be clear here. If you cannot get the right burger from an Al customer service bot, it is not an appropriate tool for education, tutoring, or any other high-stakes task.”
October 30, 2025 at 1:44 PM
mkirschenbaum.bsky.social
It’s interesting to me that none of the #criticalAI discourse (to my knowledge) has picked up Bruce Holsinger‘s novel CULPABILITY. The premise is that a self-driving car (privately owned by a family) is involved in a fatal accident—so who’s to blame? +
October 28, 2025 at 3:04 PM
lmesseri.bsky.social
Very grateful to fellow travelers who are trying to temper enthusiasm for AI, especially in the halls of the academy. @janerosenzweig.bsky.social @melhogan.bsky.social @eryk.bsky.social @histoftech.bsky.social @criticalai-journal.bsky.social @jeddobson.bsky.social @couldrynick.bsky.social and more
October 21, 2025 at 3:47 PM
nataliialaba.com
New study in Memory, Mind & Media: "Memory of the multitude and representation in AI-generated #images of #war;
#AI and #Memory special collection. doi.org/10.1017/mem....

#criticalAI @universitypress.cambridge.org
Memory of the multitude and representation in AI-generated images of war Abstract This study addresses how AI-generated images of war are changing the making of memory. Instead of asking how AI-generated images affect individual recall, we focus on how they communicate specific representations, recognising that such portrayals can cultivate particular assumptions and beliefs. Drawing on memory of the multitude, visual social semiotics, and cultivation/desensitisation theories, we analyse how visual generative AI mediates the representation of the Russia-Ukraine war. Our corpus includes 200 images of the Russia-Ukraine war generated from 23 prompts across proprietary and open-source visual generative AI systems. The findings indicate that visual generative AI tends to present a sanitised view of the war. Critical aspects, such as death, injury, and suffering of children and refugees are often excluded. Furthermore, a disproportional focus on urban areas misrepresents the full scope of the war. Visual generative AI, we argue, introduces a new dimension to memory making in that it blends documentation with speculative fiction by synthesising the multitude embedded within the visual memory of war archives, historical biases, representational limitations, and commercial risk aversion. By foregrounding the socio-technical and discursive dimensions of synthetic war content, this study contributes to an interdisciplinary dialogue on collective memory at the intersection of visual communication studies, media studies, and memory studies by providing empirical insights into how generative AI mediates the visual representation of war through human-archival-mechanistic entanglements. Keywords: visual generative AI; AI-generated images; memory of the multitude; war representation
September 23, 2025 at 7:59 AM
caddiealford.bsky.social
Check out the guest post I did for @dhandlib.bsky.social on what I read this summer, practicing the @disconetwork.bsky.social method of "technoskepticism" vis-a-vis critical AI and platformization.

dhandlib.org/recommended-...

#TeamRhetoric
#Academicsky
#CriticalAI
RECOMMENDED: What I Read This Summer: Caddie Alford ← dh+lib
dhandlib.org
September 19, 2025 at 3:23 PM
drsundar.bsky.social
🧵 7/8
Protocol: Each participant assessed 10 poems (mixed authorship). Control group made authorship judgments of AI vs Human post-evaluation.
#CriticalAI #PsychSci #poetry
January 11, 2025 at 10:33 PM
janerosenzweig.bsky.social
thanks, @cathomps.bsky.social! A few people to follow here on bsky who are very thoughtful on AI: @annamillsoer.bsky.social @criticalai-journal.bsky.social @kconrad.bsky.social @joshbrake.com @marcwatkins.bsky.social
December 31, 2023 at 7:43 PM
markwong.bsky.social
Ample evidence in #criticalAI studies has revealed #AIharms on racialised people. AI models cause harm by transmitting discrimination, toxicity, misinformation, and negative stereotypes. what is lesser known is how people makes sense of and navigate these systems and harms. 5/n
November 20, 2024 at 12:28 PM
eugeniaunivie.bsky.social
@olivia.science
@timchristiaens.bsky.social
@annanosthoff.bsky.social
@irisvanrooij.bsky.social
@elkeschwarz.bsky.social
@paulschuetze.bsky.social
@emilymbender.bsky.social
@criticalai-journal.bsky.social
@lukestark.bsky.social
@zitron.bsky.social
@eugeniaunivie.bsky.social
November 17, 2024 at 9:48 PM
annamillsoer.bsky.social
I am sympathetic to that perspective. However, I do think that "AI" is the term in use and it's possible to critique AI hype using the term. I follow @criticalai-journal.bsky.social @criticalai.bsky.social on this.
December 24, 2023 at 1:49 AM
rawx.bsky.social
"we test how OpenAI's new search feature surfaces and attributes news content", and found an error rate of 75%.

That's a damning result.

#ai #criticalai
klaudia.bsky.social klaudia jaźwińska @klaudia.bsky.social · Nov 27
Specifically, we were looking to see if it accurately returned the publisher, date and URL of the quotes we shared.
Of the 200 responses to our queries, 153 were either partially or entirely incorrect. But the chatbot only expressed a lack of confidence in its answers 7 times. (3/9)
Three examples of ChatGPT’s responses to our queries, with varying degrees of correctness Chart heading: ChatGPT's search feature was frequently wrong, but rarely unsure. Chart subheading: The Tow Center asked ChatGPT to identify the publication, date, and URL for 200 quotes. It was confidently wrong in 146 cases.
November 29, 2024 at 7:41 PM
karlfitzgerald.bsky.social
Multi national property interests are hoarding real estate data sets not to protect communities, but to pre-position capital, with #AI by its side. open.substack.com/pub/karlfitz... @tobylloyd.bsky.social @desireefields.bsky.social @criticalai-journal.bsky.social
The map was public. The tolls are private.
How privatised land data and platform finance hollow out democracy, and why CLTs are the circuit breaker
open.substack.com
October 3, 2025 at 7:30 AM
mettalrose.bsky.social
5 star launches "Make AI Mediocre Again" campaign where they flood the internet with nonsensical websites. When LLMs train on them, they learn incorrect things and spew out mediocre content so that human writers can enjoy more lesirue time. Damn! #CriticalAI #TeamRhetoric
youtu.be/73rkqkTY6dA?...
M. A. M. A. ✊ | Cadbury 5 Star
YouTube video by Cadbury 5 Star India
youtu.be
December 7, 2024 at 5:08 PM
dashishka.bsky.social
#Biden's Order to Regulate AI Further:

-> Developers of #criticalAI for public security must share safety test results and "other critical information".

-> The #NIST will create a new standards, tools and tests...

(check thread)

www.whitehouse.gov/briefing-roo...
October 31, 2023 at 10:50 AM
andrewgarton.bsky.social
Tunecore invites some musicians to "opt-in" to a AI Data Protection Program, but you also allow them use of "your content to aid your creative process or audience development, while maintaining control." The agreement does not appear nuanced enough. #CriticalAI

support.tunecore.com/hc/en-au/art...
TuneCore's AI & Data Protection Program
TuneCore’s AI & Data Protection Program is an invite-only program that allows you to identify yourself as an artist who wants to shape the future of Artificial Intelligence (AI) and Machine Lea...
support.tunecore.com
November 28, 2024 at 10:31 PM
fromtherostra.com
Dan Cryer’s “critical minimal adoption” take on GenAI in writing courses: aiminimalist.wordpress.com

Lots of ideas to consider, including “We should help administrators understand how AI creates more work for teachers, not less.”
#TeachingWriting #HigherEd #CriticalAI #EdTech
December 3, 2024 at 6:45 PM
criticalai-journal.bsky.social
#CriticalAI doesn't usually post non-AI politics; but emboldened #whitesupremacy is an indelible feature of our tech-centric times.
The "titans" behind AI hype have shown their autocratic cards. They're resentful egotists.
Let's defeat THEM along w/ Trump in November.
July 22, 2024 at 8:23 PM
jukhau.bsky.social
"Critical Topics: AI Images is an undergraduate class I created and delivered for Bradley University in Spring 2023. Now, it’s something of a beginner’s textbook for critical AI." - Eryk Salvaggio.

www.cyberneticforests.com/ai-images

#criticalAI
Critical Topics: AI Images Course — Cybernetic Forests.
An archive of an undergraduate class on the context, history and ethical issues of AI images.
www.cyberneticforests.com
February 9, 2025 at 9:51 AM
hrsnell.bsky.social
Do we have to go back entirely to physical media now? Work on offline-only computers? It's becoming increasingly difficult to trust that anything is crafted the long way by an actual human because of this AI nonsense.

#CriticalAI

www.theguardian.com/technology/2...
Meta wins AI copyright lawsuit as US judge rules against authors
Writers accused Facebook owner of breach over using books without permission to train its AI system
www.theguardian.com
June 26, 2025 at 4:11 PM
jamie2112.bsky.social
McDonald’s removes AI drive-through after order errors (BBC News) … let’s be clear here. If you cannot get the right burger from an AI customer service bot, it is not an appropriate tool for education, tutoring, or any other high-stakes task. -@CriticalAI
July 8, 2025 at 6:56 PM
rycparke.bsky.social
doaj.org/article/1ace...
@xriskology.bsky.social @criticalai-journal.bsky.social
@davetroy.com

Maoist(?) Pro-Extinctionism?
Aren't we supposedly in a Cold War with China?

Also look up Beijing Genomics Group (BGI) and the Yinfeng Life Extension Program (Cryonics)
Interplanetary Revolutions – DOAJ
The paper introduces the transhumanism and cosmic communism of Mao Zedong and discusses its relation to a Marxian conception of human nature and analogous...
doaj.org
October 23, 2025 at 9:19 PM
remyd.bsky.social
AI is taking an interpretative turn, mostly thanks to #CriticalAI and works on biases, but also because of imaging technologies and natural language processing as I argue in this piece (and like Don Ihde did on the imaging part). #NLP #Hermeneutics #PhilSky theconversation.com/lia-est-elle...
L’IA est-elle capable d’interpréter ce qu’on lui demande ?
Depuis l’avènement de ChatGPT, la machine parle notre langue. Elle doit donc interpréter nos mots, et nous sa réponse. Comment analyser philosophiquement cette « interprétation artificielle » ?
theconversation.com
June 4, 2024 at 11:22 AM
mkmayhall.bsky.social
So timely for me. Just finished a unit on the ethics of AI w/Ss based on
@torreytrust.bsky.social ‘s open access slideshow & they were appalled at what they’re not being taught. Genuinely outraged. We need more critical AI voices like @criticalai-journal.bsky.social & @eryk.bsky.social as well.
January 31, 2025 at 3:31 PM
criticalai-journal.bsky.social
#CriticalAI Editorial Collective member Christopher Newfield has a great blog post out on the crisis in creating and disseminating knowledge.

Research funding is migrating away from the basics + from much-needed social and cultural knowledge.
isrf.org/2024/04/03/t...
April 11, 2024 at 6:40 PM