Lightnews — Scholar-powered news

Reposted by Eugenia Stamboliev

Roland Meyer @bildoperationen.bsky.social · 1d

This is both fascinating & frightening. There is a demand for images that show and confirm what people already know and believe to be true, and AI slop meets that demand—with visual content that emblematically visualizes what the headlines report. One could call it synthetic hypervisualization
1/

Matt Novak @paleofuture.bsky.social · 1d

That TikTok account has a few fake videos of reporters getting detained. They make these videos because it’s a big story in the news. Real reporters are getting arrested. And so people are making AI videos to get engagement on TikTok.

dontbeonline • 1d ago
Journalist gets detained while doing her job #ForYou, #viral, #foryoupage, #tikto... more
• Contains: Bad Boys (The...

9 120 290

Eugenia Stamboliev @eugeniaunivie.bsky.social · 3d

‚Some researchers see this behaviour as a serious threat, whereas others call it hype. So should these episodes really cause alarm, or is it foolish to treat LLMs as malevolent masterminds?‘ I am torn here 🫤 a/ legally, murder plotting requires intent&accountability
www.nature.com/articles/d41...

AI models that lie, cheat and plot murder: how dangerous are LLMs really?

Tests of large language models reveal that they can behave in deceptive and potentially harmful ways. What does this mean for the future?

www.nature.com

Eugenia Stamboliev @eugeniaunivie.bsky.social · 12d

Lots of things wrong here, from the maybe realness to the only toolness (?). I don’t get why fake women equally have to embody fake beauty standards, which real women have been tortured with forever a/ now, get resold as standardised womenness.
#syntethicdiscrimination

variety.com/2025/film/gl...

AI Actress Tilly Norwood Draws Backlash From Melissa Barrera, Lukas Gage and More Hollywood Names as Creator Defends Her as a ‘New Tool’ and ‘Not a Replacement for a Human Being’

Tilly Norwood's creator has responded to the angry backlash against the AI actress, saying she's 'not a replacement for a human being'

variety.com

1 2

Eugenia Stamboliev @eugeniaunivie.bsky.social · 21d

Surprise surprise, hm, not really. www.computerworld.com/article/4059...

OpenAI admits AI hallucinations are mathematically inevitable, not just engineering flaws

In a landmark study, OpenAI researchers reveal that large language models will always produce plausible but false outputs, even with perfect data, due to fundamental statistical and computational limi...

www.computerworld.com

4 8

Eugenia Stamboliev @eugeniaunivie.bsky.social · Aug 24

#CreativeSurveillanceHacks
Love the (analogue) #hacking twist, hate that we need this…
conemagazine.com/culture/capa...

“Opacity Is a Right”: How Cap_able Uses Fashion to Fool Facial Recognition Technology - CONE Magazine

Adversarial fashion by Cap_able shields identities from facial recognition and AI surveillance, merging privacy with style.

conemagazine.com

Eugenia Stamboliev @eugeniaunivie.bsky.social · Aug 22

Well I don’t want to, but more importantly, I also cannot access my files. A shame …

2

Reposted by Eugenia Stamboliev

Barbara Prainsack @bprainsack.bsky.social · Aug 19

everyone who believes that AI will fix democracy: Read this piece by @himself.bsky.social‬ ‪@hahrie.bsky.social‬

knightcolumbia.org/content/ai-a...

AI and Democratic Publics

knightcolumbia.org

8 16

Reposted by Eugenia Stamboliev

Keith Wooo-schmann Wilson 👻 @keithwilson.eu · Aug 18

The FDA is using an LLM to speed up drug approvals. 🤦‍♂️ #AIEthics #Trumpland #MakeAmericaSickAgain

Victor Galaz @vgalaz.bsky.social · Aug 18

”Yet behind the scenes, the agency’s slick AI project has been greeted with a shrug — or outright alarm.”

edition.cnn.com/2025/07/23/p...

FDA’s artificial intelligence is supposed to revolutionize drug approvals. It’s making up nonexistent studies. | CNN Politics

Insiders tell CNN the FDA’s AI is “hallucinating” studies and can’t access key documents. Agency leaders insist the AI is getting better, and use is not mandatory.

edition.cnn.com

1 4 3

Eugenia Stamboliev @eugeniaunivie.bsky.social · Aug 17

www.derstandard.at/story/300000...

Ihr wollt eine gerechte Stadt? Macht sie feministisch!

Warum eine feministische Stadtplanung allen nützt – und wie wir sie erreichen

www.derstandard.at

1

Eugenia Stamboliev @eugeniaunivie.bsky.social · Aug 15

Der Song zum Kopf 🫠 solidarity💪🏼

1

Eugenia Stamboliev @eugeniaunivie.bsky.social · Aug 3

😳

Eugenia Stamboliev @eugeniaunivie.bsky.social · Aug 3

Can‘t wait for ‚…however, he could’ve reached his career goals faster and according to his strategic outlook. Unfortunately, he showed little management talent throughout this employment at xyz…‘ ‚In conclusion, … he was a nice chap who was paramount to his family tree‘ … the future looks bleak 😒

1

Eugenia Stamboliev @eugeniaunivie.bsky.social · Aug 1

I wrote something on #syntheticdiscrimination; a form that is rendering existing women invisible online due to the hypervisible, synthetic female imagery made by GenAI and taking over.

Publication👇🏻 (for now, only in German, @unesco.org)
www.unesco.at/fileadmin/us...

1

Reposted by Eugenia Stamboliev

Roland Meyer @bildoperationen.bsky.social · Jul 28

Great piece on #slopaganda: «Whether it's a self-assured small social media account … or, indeed, a department of the US government, the viewing public isn't the only target of this new form of propaganda. In this manner, the creator also experiences the same reassurance from interactions online.»

🔫Scott Martian Invaders🛼 @youcaughtscott.com · Jul 9

This week I wrote about the unstoppable tide of "Slopaganda." Traditional propaganda was not meant to convince, but to assure the public that prior beliefs were popular and accepted. What happens when you allow anyone to create propaganda at the stroke of a keyboard? A supercharged feedback loop.

Journalism in the Age of Slopaganda

How does one stem the tide of disinformation?

readthecatch.ca

24 53

Eugenia Stamboliev @eugeniaunivie.bsky.social · Jul 27

#uefa2025 #frauenem #orf nach nur 5 Minuten kommentiert der Sprecher, dass #jennihermoso leider nicht dabei ist ‚die die damals in den Kuss Skandal verwickelt war‘. Nope, nix da. Sie war nicht so verwickelt, sondern wurde von LB belästigt. Solange wir so sprechen, ist alles nur kleine Lappalie.

1

Eugenia Stamboliev @eugeniaunivie.bsky.social · Jul 8

‘Power always needs abstraction. To be controlled, reality’s unwieldy complexity needs to be slotted in models, …that allow for standardization and manipulation. Nature and social life are bureaucratically indigestible in their raw form and must be pre-processed to be seen and shaped.’

Noema Magazine @noemamag.com · Jul 1

As we transition from industrial to digital modernity, business tycoons no longer view humans as cogs in machines. Now, as platforms become all-powerful, a new, stealthier form of domination is emerging, @pettertornberg.com & @uitermark.bsky.social argue.

The Ascendance Of Algorithmic Tyranny | NOEMA

As today’s platforms become all-powerful, the metaphors we use to describe our digitally infused world exemplify a new, stealthier form of domination that is emerging.

www.noemamag.com

1 2

Eugenia Stamboliev @eugeniaunivie.bsky.social · Jun 14

Despite darkness, anger or sadness, may there be a rainbow ahead.

2

Eugenia Stamboliev @eugeniaunivie.bsky.social · Jun 11

I am guiding/giving/performing a workshop on #AIethics (my twist), #art (my non-interpretation) and #AIslob (you know, Jesus as shrimp etc). This Saturday afternoon, 14.6, in #Vienna, join if you want 👇🏻🌈

(No prompting, no genius narratives, maybe crayons 🖍️)

www.semmelweisklinik.at/collective-r...

Eugenia Stamboliev @eugeniaunivie.bsky.social · Jun 8

‚It would be a disaster for our students to have no idea how to effectively use one of the most powerful tools that humanity has ever created,” Brown said. “AI is such a powerful tool for self-education…‘ To me the disaster is utter worldly incompetence and climate collapse, but that’s just me 🤨

Hypervisible @hypervisible.blacksky.app · Jun 8

🤷🏿‍♂️

Brown said AI is here to stay, so banning it is “shortsighted.” He encouraged students to have discussions about ethics and philosophy with AI chatbots, asked them to write papers using AI however they’d like and used AI to help create dialogues between two sides of a controversial topic to demonstrate educated arguments on both sides.

1

Eugenia Stamboliev @eugeniaunivie.bsky.social · Jun 5

Ich bin wieder mal kurz im Radio, es geht um #KI und Uni und ja, noch macht es einfach mehr Spaß mit Menschen zu sprechen @fm4.orf.at
sound.orf.at/podcast/fm4/...

FM4 Auf Laut

Was brennt uns unter den Fingernägeln? Was regt auf oder an und welches Thema hat es verdient, ein größeres Forum zu bekommen? Worüber man nicht schweigen kann, darüber muss man diskutieren! Jede Woc...

sound.orf.at

4

Eugenia Stamboliev @eugeniaunivie.bsky.social · Jun 1

Don’t want to dish on my former homies (yay Devon peeps!), but let’s face is, there ain’t no Turing test for this genius communication loop…

1

Eugenia Stamboliev @eugeniaunivie.bsky.social · May 23

www.bloodinthemachine.com/p/dismantlin... 🔥 #empireofai @bcmerchant.bsky.social @karenhao.bsky.social

Dismantling the Empire of AI with Karen Hao

An interview with one of the best working AI journalists on her new book, why AI companies act like empires, and how to strike back.

www.bloodinthemachine.com

3

Eugenia Stamboliev @eugeniaunivie.bsky.social · May 11

I just learned about the depth of these Thiel’ian approaches and tentacles - thanks to Saturday’s talk by @cryptologocentric.bsky.social … and here we go, now it’s Financial Times worthy. I am confused about who’s doing critique these days…
www.ft.com/content/d332...

How a little-known French literary critic became a bellwether for the US right

René Girard is best known for his theory of ‘mimetic desire’. Now Peter Thiel and the vice-president are among his fans

www.ft.com

2

Eugenia Stamboliev @eugeniaunivie.bsky.social · May 9

⚡️⚡️⚡️

The AI Boom Is Draining Water From the Areas That Need It Most

A Bloomberg News analysis finds that roughly two thirds of new data centers built or in development in the US since 2022 are in places with high levels of water stress.

www.bloomberg.com

10 12

Eugenia Stamboliev @eugeniaunivie.bsky.social · May 8

Not sure Jesus would approve these kinds of blasphemic interventions, but well …
www.science.org/content/arti...

AI predicted the next pope. Did it get it right?

Researchers use algorithms to study political factions within the Catholic Church

www.science.org