csmapnyu.org
We work to strengthen democracy by conducting rigorous research, advancing evidence-based public policy, and training the next generation of scholars.
https://csmapnyu.org/links
384 posts
9,780 followers
617 following
Regular Contributor
Active Commenter
comment in response to
post
Here's the full piece from @zevesanderson.com @solmg.bsky.social & Scott Babwah Brennen
comment in response to
post
If TikTok takes these measures seriously, they not only may ameliorate concerns from policymakers in the U.S. and abroad but also could establish a voluntary transparency model that supports more rigorous research, better policymaking, and healthier platforms.
comment in response to
post
There's no evidence that China acted on Russia's behalf. Instead, the EU is investigating potential “automated exploitation” of the algorithmic feed by Russian actors. This highlights how any algorithmic platform can be abused by adversarial actors, whether they're controlled by China or not.
comment in response to
post
In the end, even if TikTok is sold to a non-Chinese buyer, the threat of foreign influence remains. Just a few months ago a Romanian high court canceled an election after a Russian-sponsored propaganda campaign on TikTok led to a surprise victory for a far-right candidate.
comment in response to
post
Long term, TikTok can commit to more robust data-sharing approaches, including establishing data clean rooms for sensitive data, engaging in academic collaborations to allow for on-platform experimentation, and committing to not taking legal action against researchers who scrape data.
comment in response to
post
While significant attention has been (and should be) paid to how policymakers can improve transparency, TikTok itself can make changes. To start, it can improve its researcher API, the guidelines of which run counter to academic norms around data retention, publication, and replication.
comment in response to
post
It's likely too late for TikTok to make meaningful changes to Project Texas to allay the govt's national security concerns. But as the Trump admin considers how to save TikTok for American users, mechanisms to create data access for independent researchers should be a crucial part of the discussion.
comment in response to
post
What was the net effect? With access to the algorithm but w/o access to large-scale platform data, analysts know far less about Twitter/X than they did before. This emphasizes that while transparency is important to identify foreign influence, the right approach to transparency is equally important.
comment in response to
post
Twitter/X is a useful case study. Before Musk took over, Twitter had robust data-sharing practices, allowing independent researchers to study a range of topics, including foreign influence. Musk shut down data access and publicly released the Twitter's source code, which provided little insight.
comment in response to
post
But analyzing the source code wouldn't actually tell us much. Recommendation algorithms are complex & dynamic systems that leverage user data to determine what shows up in our feeds. To measure how the algorithm delivers content, we have to actually observe users' feeds.
comment in response to
post
In passing the divestment order, lawmakers feared the Chinese govt could access US user data or push propaganda into American users' feeds. TikTok tried (unsuccessfully) to address these concerns via Project Texas, which allowed Oracle to inspect TikTok's source code to look for foreign influence.
comment in response to
post
A helpful summary from @markcoddington.bsky.social and @sethclewis.bsky.social latest newsletter
rq1.substack.com/p/how-gender...
comment in response to
post
And here's the thread we posted at the time with more detail.
comment in response to
post
You can read a summary of both papers on our website from back in November.
comment in response to
post
This is the second study from this project. The first, published in November at PNAS Nexus, found that Latinos who use Spanish-language social media for news were 11-20 % points more likely to believe false political narratives than those w/English-language content.
comment in response to
post
3) Latinos who rely on Spanish-language news on social media (relative to those who use English news) are more likely to believe in election fraud narratives about the 2020 presidential election.
comment in response to
post
2) Latinos were more likely than whites (53 percent to 32 percent) to turn to social media for information about Covid-19.
comment in response to
post
The research yielded three main findings:
1) Latino online political activity, such as sharing news and discussing politics, is on par with white Americans, with one notable exception: Latinos were much more likely to use WhatsApp than whites (57% to 15%).
comment in response to
post
This paper is based on a survey examining social media political activity of 2,326 Latinos & 769 non-Hispanic whites. We augmented the survey with digital trace data from a subset of respondents.
comment in response to
post
Considerable research examines how Americans' online media consumption relates to political behavior. Much less is known about Latino online media consumption. CSMaP launched a research project to study this landscape around the 2022 midterms.
comment in response to
post
There’s a pretense of empiricism in Zuck announcement. In reality, Notes is mixed. Fact-checking impact unclear. Our ability to measure effects ⬇️ as FB & X eliminated transparency tools. Past studies showed no anti-conservative bias.
The new policies all pass the buck to users. That’s the story.
comment in response to
post
So, two questions to monitor:
1) How will the signal from the new “community notes” system Meta puts in place differ from the existing third-party fact checking one?
2) Does today’s announcement suggest forthcoming changes in how Meta intends to utilize that signal in the feed algorithm?
comment in response to
post
We know from prior research that Meta can increase and decrease the relative weight of that signal depending on external circumstances, but removing it entirely would likely significantly change the nature of the platform.