🧵on my new paper "Synthetic personas distort the structure of human belief systems" w Roberto Cerina I'm v excited about...
🚨 Do synthetic samples look like human samples?
We compare 28 LLMs to the 2024 General Social Survey (GSS) to find out + develop host of diagnostics...
Leaders from industry, startups, government, & nonprofits joined us for a "What the Heck Does A Data Scientist Do?" panel.
Panelists:
⭐CDS Prof Foster Provost
⭐Goguma CTO Haftan Eckholdt
⭐CDS alums @swapneel.bsky.social & Tatenda Ndambakuwa
⭐NYC Campaign Finance Board's Sarath Kareti
A huge thanks to the wonderful @justinhendrix.bsky.social for supporting, guiding, and helping us pull this together!!!
Among others, it includes lessons on:
- trustworthy AI for journalism with @swapneel.bsky.social,
- tips on safeguarding mental health in the age of digital disinfo with @jospang.bsky.social,
- and insights on tackling disinformation from the Global South with @hwasser.bsky.social.
Learning resource: Have you checked out the videos for our #TacklingDisinformation course?
Our playlist over on YouTube gives you an overview of current concepts and discussions.
www.youtube.com/playlist?lis...
Our work presents an overview of the historical attempts, successes, and subsequent challenges that emerged that dictated the long term impact of regulation on technological and socioeconomic progress in healthcare, aviation, finance, and automobiles. Read the paper for more:
Regulation is framed as adversarial to progress by those it subjects to accountability for their actions, especially by big tech, social media, and AI companies. It's more important than ever before to get things right and balance between consumer protection and government overreach.
Other sectors including healthcare, finance, and aviation have had similar concerns around transparency and decades of regulatory attempts have resulted in solutions that -- while imperfect -- work to protect consumers against harm, and to audit providers within these sectors for accountability.
Really excited to publish an article with Sam and Deepika discussing concrete steps for transparency regulation for digital platforms!
Social networks aren't the first case of incredibly powerful, incredibly opaque systems that cater to millions of users across the world.
And thanks to an AI assist, we organized these so you can parse through some of these platform interventions and organize others you see in practice!
interventions-report.vercel.app
Take a look at our paper and let us know if you have thoughts on how to improve coverage of platform interventions in the wild!
papers.ssrn.com/sol3/papers....
We organize interventions by their focus, scope, driver, and user journey, and unify the language used to describe them. For practitioners, we provide a clear example of how this helps to discuss interventions deployed across TikTok, YouTube, and Instagram.
Prior work has attempted to organize interventions in different ways which is great, but it is still challenging to consider the relative impact of *actual* platform interventions since the focus has been limited to thinking along academic terms so we shift gears towards practical implications.
Our work on the design of social media interventions will be presented at TPRC 2025! We discuss the need for common language to discuss disjoint platform interventions and organize them into classes of interventions commonly applied by platforms so we can assess their collective impact and trends.
...that were providing a free service monitoring digital harms from social media data access have had to wait ages for any semblance of meaningful support to emerge for regaining comparable data access from digital platforms.
I get that business and academia aren't the same but there is some irony in the immediate furore created around making crypto and fintech startups pay for data access and "cripple" their business models, while academics and researchers ...
Last year, we published a paper showing that AI models can "debunk" conspiracy theories via personalized conversations. That paper raised a major question: WHY are the human<>AI convos so effective? In a new working paper, we have some answers.
TLDR: facts
osf.io/preprints/ps...
I usually add a disclaimer that my personal experiences have been overwhelmingly positive but that's not always been the case of my peers, plus there's always a chance things go sideways with a shaky federal funding landscape so this advice is additionally timely for those coming in now.
I want to say most advisors are good people but there are always a few bad apples where the vibes are off.
Ensure that you are intentional about gathering the experiences of past mentees as you go into the acceptance process for graduate school offers. All the best!
When you're going into your PhD you want to consider what kind of person your advisor has been to past mentees especially thinking of the doors they have opened and the rooms they have been in.
Consider whether you'd like to be in the same rooms and part of the same communities.
A mentee told me this should have been a social media post so here it is.
A great advisor opens doors for you that you never even imagined you would be good enough for.
A good advisor opens doors for you to rooms they've been in.
A bad advisor uses your name to walk into the room themselves.
Follow the program and agenda at the Policy Center for the New South website with the program and session recordings available online! Apply to the Atlantic Dialogues Emerging Leaders #ADEL program for their next cycle later this year, if you work in this space.
www.policycenter.ma/adel-communi...
The world is moving fast and it’s difficult to keep up! Ask yourself how often you update your own perspective by making space for others to challenge you—especially others from your own teams. Set an example for others to follow!
There’s no #technology that guarantees you truthful answers 100% of the time because even such technology relies on third parties and the internet to draw its inference.
Allow for uncertainty in conversations so that the public starts to acknowledge knowledge limitations and rebuild trust in democratic institutions instead of getting carried away by ideologues and misinformation. Facts almost always evolve and what is true today may not remain true tomorrow.
youth panel at atlantic dialogues with emerging leaders
Let’s be more accepting of our own limitations and make space for those who may contribute alternate expertise, especially those from the younger generation who have demonstrated outsized potential in their field.
It was an honor to represent the 41 emerging leaders selected from over 20 nationalities in the Atlantic Dialogues program in December. A tl;dr on my comments to the dignitaries and political leaders from Africa, Asia, and Europe about the "Wider Challenges in the Atlantic Basin" 👇🧵
New year, new updates to my team's work on #ai, #digitaltrust, #misinformation, and platform governance up on mehtaver.se and hopefully more writing on the blog this year! Reach out if any of these topics are of interest and I'm happy to chat about our plans!
Totally fascinating replies from @aoc.bsky.social/Trump split ticket voters on her IG.
Proud to share the Integrity Institute's second election integrity guide for online platforms (adding to our first guide earlier this year)! @katieharbath.bsky.social and I, and the fantastic group of integrity professionals at the Institute, hope people find this helpful heading into 2024 elections
