Mor Naaman
@informor.bsky.social
Cornell Tech professor (information science, AI-mediated Communication, trustworthiness of our information ecosystem). New York City. Taller in person. Opinions my own.
Here it is, the most damning academic figure in the history of all figures: Russian IRA accounts deeply embedded in both sides of the 2016 Twitter BLM "discussion" (by @katestarbird.bsky.social & team, 2018).
THIS is what our information ecosystem supports. Twitter just made it 1% more visible.
THIS is what our information ecosystem supports. Twitter just made it 1% more visible.
November 25, 2025 at 1:30 PM
Here it is, the most damning academic figure in the history of all figures: Russian IRA accounts deeply embedded in both sides of the 2016 Twitter BLM "discussion" (by @katestarbird.bsky.social & team, 2018).
THIS is what our information ecosystem supports. Twitter just made it 1% more visible.
THIS is what our information ecosystem supports. Twitter just made it 1% more visible.
I don't know if @jilltxt.bsky.social does but I just heard her brilliant talk documenting what sounded like a similar problem
November 25, 2025 at 1:40 AM
I don't know if @jilltxt.bsky.social does but I just heard her brilliant talk documenting what sounded like a similar problem
No but I can make it so they are uploaded less
November 25, 2025 at 1:34 AM
No but I can make it so they are uploaded less
To be clear for other readers, this was a story about independence, not perks
November 24, 2025 at 9:53 PM
To be clear for other readers, this was a story about independence, not perks
TVs these days are too thin for rabbit ears! 🐰😱
November 23, 2025 at 11:17 PM
TVs these days are too thin for rabbit ears! 🐰😱
Haha yes it is not perfect
November 23, 2025 at 3:14 PM
Haha yes it is not perfect
In summary, the moment captured here explains why I am now a professor.
November 23, 2025 at 2:18 PM
In summary, the moment captured here explains why I am now a professor.
The Missing Kayaker
What happened to Ryan Borgwardt?
www.theatlantic.com
November 22, 2025 at 3:01 PM
Yep, that paper makes it clear 😭
November 22, 2025 at 3:23 AM
Yep, that paper makes it clear 😭
What do you mean "manually reviewed" participants? How did you detect AI? And was it 30/150?!
November 22, 2025 at 12:26 AM
What do you mean "manually reviewed" participants? How did you detect AI? And was it 30/150?!
Now here's something to be alarmed about @mims.bsky.social -- even if social scientist do not turn to these tools, these tools (simpler, revenue-maximizing ones) will find them.
bsky.app/profile/did:...
bsky.app/profile/did:...
Summary: a (rather simple) AI bot can not only successfully evade survey-level protections, it can also mimic personas accurately across questions, parse photos/videos, and infer the researcher’s hypotheses to produce data that confirms them.
The official paper link:
www.pnas.org/doi/10.1073/...
The official paper link:
www.pnas.org/doi/10.1073/...
November 21, 2025 at 3:55 PM
Now here's something to be alarmed about @mims.bsky.social -- even if social scientist do not turn to these tools, these tools (simpler, revenue-maximizing ones) will find them.
bsky.app/profile/did:...
bsky.app/profile/did:...
One example: the bot performed a prescribed "education level" accurately across multiple "enter state capital" questions -- though it does show some vulnerability, if you examine the rate of "own state correct". Phew!
(but, surveys without additional protections are now officially a fool's errand)
(but, surveys without additional protections are now officially a fool's errand)
November 21, 2025 at 3:17 PM
One example: the bot performed a prescribed "education level" accurately across multiple "enter state capital" questions -- though it does show some vulnerability, if you examine the rate of "own state correct". Phew!
(but, surveys without additional protections are now officially a fool's errand)
(but, surveys without additional protections are now officially a fool's errand)
