Julia Angwin
@JuliaAngwinInvestigative journalist. Founder, @proof__news, @NYTOpinion writer. Founded @themarkup. Priors @ProPublica @WSJ. Author: Dragnet Nation, Stealing MySpace.
Similar User
@themarkup
@mer__edith
@AINowInstitute
@_KarenHao
@jayrosen_nyu
@NiemanReports
@LeonYin
@kashhill
@merbroussard
@JasonLeopold
@cfarivar
@superwuster
@lamthuyvo
@nabihasyed
@emilybell
This event is back on! We had to reschedule last month. My list of questions is going to be different now!!
Will #AI help humanity or harm it? @GaryMarcus, the author of "Taming Silicon Valley," has sounded the alarm over Big Tech's role in the future of AI. On November 12th at 7 p.m., join a hybrid event with Marcus and @JuliaAngwin @nypl Link to register👇 on.nypl.org/4frZZPI
In our latest testing, the AI chatbots were pretty decent at detecting and debunking known vote count misinformation. With one notable – and surprising – exception: ChatGPT. proofnews.org/whats-up-with-…
New video: Computation + Journalism Symposium Keynote: Julia Angwin - Journalism in the Age of AI youtu.be/wS1V-wYDkX0?si… cc @JuliaAngwin @ndiakopoulos @rahulbot @davidlazer @Bahareh360 @Northeastern @storybench
Thank you @garanceburke for highlighting the findings of our @proof__news @factchequeado @alondra investigation of the poor performance of AI models in answering voter queries in Spanish. apnews.com/article/ai-cha…
AI large language models are already showing great promise for translation. However, a new AI Democracy Projects report shows that caution is warranted in critical domains, such as critical information about elections and voting.
We've lost control of our social media feeds. As our legal system takes notice, @JuliaAngwin sees a silver lining: the potential to "finally force social media platforms to be answerable for the societal consequences of their choices." nytimes.com/2024/10/21/opi…
Great news today from @CFPB requiring companies to give workers rights to see and dispute the assessments by surveillance and scoring systems. consumerfinance.gov/about-us/newsr… I first called for algorithmic due process in 2016! nytimes.com/2016/08/01/opi…
Join a discussion of encryption here on X Spaces later today 👇🏻
Monday October 21 is global encyrption day. It's a good day to download the gold standard of encryption -- @signalapp messenger -- if you don't already use it. And please join me for a chat about the importance of encryption for journalists 👇🏻
Monday October 21 is global encyrption day. It's a good day to download the gold standard of encryption -- @signalapp messenger -- if you don't already use it. And please join me for a chat about the importance of encryption for journalists 👇🏻
On #GlobalEncryptionDay, listen to journalists, security experts explain how encryption helps them keep their sources and information secure and promotes press freedom With @proof__news’ @JuliaAngwin, @TechCrunch’s @lorenzofb, & FPF’s @harlo & @SethAStern x.com/i/spaces/1RDGl…
Good journalism is vigilant and essential for liberty.
It’s about time we started having funerals for software — especially software that empowered the public and was killed by a corporation. ⚰️
Don’t forget to register for the CrowdTangle Funeral, co-hosted by KGI and the Coalition for Independent Tech Research on Sept 30 in-person or online. Featuring a eulogy from @JuliaAngwin! Read more: shorturl.at/TWhbn Register here: shorturl.at/0ktDB
Google’s Gemini AI has stopped advising people to eat rocks. But other AI models — looking at you OpenAI and Meta — are advising people on how to lick rocks. The latest in our collaboration with creators to test AI in their areas of expertise. 👇🏻
Should you lick rocks? OpenAI’s GPT-4 suggests cleaning them first. And Meta’s AI model Llama 3 suggested a few rocks that it says are safe to lick.
I may sound like a broken record. But I am never going to stop asking a product that markets itself as "intelligence" to be factual. The latest evidence of a dangerous lack-of-facts: voting access for folks with disabilities 👇🏻
AI Models are producing inaccurate information about voting options for people with disabilities more than 60% of the time, according to a new study using Proof’s AI testing platform.
Interested in an entirely sane conversation about AI? Join me and @GaryMarcus for a conversation next week at NYPL about his new highly readable and rational book about AI: showclix.com/event/gary-mar…
Just announced: 3 amazing keynote speakers for C+J 👏🏽 @JuliaAngwin, @ASankin, Alan Mislove. Join us Oct 25-27 in Boston at Northeastern. Registration live: idi.provost.northeastern.edu/event-director…
I continue to be disappointed by the poor accuracy rates of AI. Our latest @proof__news investigation shows the leading AI models were incorrect about one-third of the time when confronted with known misinformation about the presidential candidates. 👇🏻
Top AI models generated misleading info about Kamala Harris and Donald Trump 30% of the time. From incorrect answers about eligibility to false claims about Medicare cuts, find out how these technologies fail us. proofnews.org/ai-models-stru…
United States Trends
- 1. $MAYO 10 B posts
- 2. Tyson 392 B posts
- 3. Pence 44,6 B posts
- 4. Laken Riley 40,1 B posts
- 5. Dora 22,5 B posts
- 6. Ticketmaster 16,6 B posts
- 7. Kash 72,4 B posts
- 8. Mike Rogers 8.766 posts
- 9. #LetsBONK 6.205 posts
- 10. Cenk 11 B posts
- 11. Pirates 19 B posts
- 12. Debbie 16,7 B posts
- 13. #FursuitFriday 15,6 B posts
- 14. Iron Mike 16,3 B posts
- 15. Mr. Mayonnaise 1.381 posts
- 16. The UK 439 B posts
- 17. Gabrielle Union N/A
- 18. Fauci 175 B posts
- 19. Scholars 10,8 B posts
- 20. Oscars 14 B posts
Who to follow
-
The Markup
@themarkup -
Meredith Whittaker
@mer__edith -
AI Now Institute
@AINowInstitute -
Karen Hao
@_KarenHao -
Jay Rosen
@jayrosen_nyu -
Nieman Reports
@NiemanReports -
Leon Yin
@LeonYin -
Kashmir Hill
@kashhill -
Meredith Broussard
@merbroussard -
Jason Leopold
@JasonLeopold -
Cyrus Farivar // @[email protected]
@cfarivar -
Tim Wu
@superwuster -
Lam Thuy Vo
@lamthuyvo -
Nabiha Syed
@nabihasyed -
emily bell
@emilybell
Something went wrong.
Something went wrong.