iiinova — where your digital self keeps evolving

Reposted by Eve Fleisig @ FAccT 2026

Alison Gopnik @alisongopnik.bsky.social · 01/09/2026

Thanks, but actually my favorite line didn't make it to the piece. Does Odysseus feel guilty about the Trojan war? Lots of discourse on this question even though he doesn't exist! Is he conscious? AI agents are fictive characters abstracted from text - like Odysseus.

28918

Reposted by Eve Fleisig @ FAccT 2026

Prof. Nicole Holliday @mixedlinguist.bsky.social · 19/08/2026

In a keynote at Sociolinguistics and AI Conference, Prof. Rodney Jones analyzes discourses about “consciousness” and the “great awakening of AI” and explains ideologies of prompts as incantations. This is technoanimism, and it’s more evidence that people are orienting towards “AI” as religion.

“Waking up AI is about waking you up” TikTok clips

25314

Eve Fleisig @ FAccT 2026 @efleisig.bsky.social · 11/07/2026

Congrats, a great paper and so well deserved!!

030

Reposted by Eve Fleisig @ FAccT 2026

M.J. Crockett @mjcrockett.bsky.social · 05/07/2026

Joseph Weizenbaum's 1976 book, Computer Power and Human Reason, has long been out of print - used copies sell for hundreds of dollars. It's wild that this is not more widely available given how much his ideas still apply to the world we're in now. 🧵/

15348122

Reposted by Eve Fleisig @ FAccT 2026

Lauren Chambers @laurenmarietta.bsky.social · 26/06/2026

“This is the ghost at the center of our field: we keep being right, but wrong keeps winning. Being right is not the same as being enough.” #FAccT2026

0114

Eve Fleisig @ FAccT 2026 @efleisig.bsky.social · 25/06/2026

Genevieve Smith and I are presenting on “Standard Language Ideology in AI-Generated Language” today at 10:45 in Ballroom East!

000

Eve Fleisig @ FAccT 2026 @efleisig.bsky.social · 25/06/2026

🇨🇦I’ll be at #FAccT2026! If you're interested in... - Societal impacts of LLMs - Improving validity of NLP evaluations - Preference learning when users disagree - Effects of personalized models - Preserving users' agency + skills in human-AI interaction ...let's chat, DMs open!

010

Reposted by Eve Fleisig @ FAccT 2026

Prof. Nicole Holliday @mixedlinguist.bsky.social · 17/06/2026

Check out this thread about some recent research by brilliant Berkeley students on how listeners attribute social identities to synthesized voices! Full paper: www.degruyterbrill.com/document/doi...

degruyterbrill.com

1228

Eve Fleisig @ FAccT 2026 @efleisig.bsky.social · 17/06/2026

📄 now out in Phonetica: degruyterbrill.com/document/doi... Many thanks to my collaborators: Julian Vargo, Niko Schwarz, Veronica Grajeda, Abigail Roberts, Rhosean Asmah & Nicole Holliday @mixedlinguist.bsky.social. Loved combining NLP technical tools with phonetic analysis!

degruyterbrill.com

092

Eve Fleisig @ FAccT 2026 @efleisig.bsky.social · 17/06/2026

In other words: listeners hear supposedly "neutral" voices, and still assign them full demographic personas based on how they sound. Then, software engineers use these voices in real applications that match stereotypes of those personas, perpetuating the stereotype.

151

Eve Fleisig @ FAccT 2026 @efleisig.bsky.social · 17/06/2026

Linguists have long known that people use vocal features like pitch & breathiness to index identity, and will judge others' identities on voice alone. So...if a synthesized voice seems to index an identity, people will perceive it that way—and stereotype it that way, too—no matter how it's marketed.

151

Eve Fleisig @ FAccT 2026 @efleisig.bsky.social · 17/06/2026

People are remarkably consistent about assigning demographic and personality traits to these voices, even when they *know* they're AI-generated. They assign full personas to the voices just based on how they sound, with identities and stereotypes to boot.

161

Eve Fleisig @ FAccT 2026 @efleisig.bsky.social · 17/06/2026

Low-pitched masculine voices get rated as more likely to be Black or Hispanic, and get used for voices of inanimate objects, like smart devices. Voices perceived as younger get better ratings across nearly every axis.

161

Eve Fleisig @ FAccT 2026 @efleisig.bsky.social · 17/06/2026

Take "coral", a relatively low-pitched, non-breathy voice. Listeners rated it as sounding like an older woman. The result? It gets rated as less competent, less trustworthy, and less friendly. In real code, it's used for formal work/business applications.

141

Eve Fleisig @ FAccT 2026 @efleisig.bsky.social · 17/06/2026

We traced how these AI voices perpetuate biases, start to end: how they sound (acoustic measurements) -> how people perceive them (listener ratings) -> the applications they get used for (GitHub repo analysis).

161

Eve Fleisig @ FAccT 2026 @efleisig.bsky.social · 17/06/2026

Remember the backlash against all AI voices being feminine—Siri, Alexa, etc.? Now, companies use a wider range of AI voices with ambiguous names like “alloy” and “shimmer.” Bias solved forever—ha, no. What happens when people hear these supposedly neutral voices?👇

1105

Reposted by Eve Fleisig @ FAccT 2026

Melanie Mitchell @melaniemitchell.bsky.social · 17/06/2026

@author-calnewport.bsky.social does a fantastic job writing about AI, and this one is his best yet. Gift link in quoted post.

0217