Aarushi Kalra

@chidiya.bsky.social

Postodoctoral Prize Research Fellow at Nuffield and Economics, Oxford || Digital Economies/ Development/ Political Economy || Cofounder Bahujan Economists || https://aarushirita.github.io/

In case somebody missed this yesterday, while watching a political car-crash unfold: "The Means of Prediction - How AI Really Works (and Who Benefits)" is now in the UChicago Press catalog, and available for pre-order online! press.uchicago.edu/ucp/books/bo...

The Means of Prediction

An eye-opening examination of how power—not technology—will define life with AI. AI is inescapable, from its mundane uses online to its increasingly consequential decision-making in courtrooms,…

press.uchicago.edu

Second: Meta seriously weakened its hate speech standards in a way that makes all protected groups generally, and immigrants, women, and LGBTQ+ people specifically, much more vulnerable to abuse on the site. 🧵 5/11

Here's a our new paper "The Consumer Welfare Effects of Online Ads: Evidence from a 9-Year Experiment" We analyze a long-running field experiment of online advertising. We find no significant differences in valuations between users who see ads vs. those who don't. www.nber.org/papers/w32846

The Consumer Welfare Effects of Online Ads: Evidence from a 9-Year Experiment

Founded in 1920, the NBER is a private, non-profit, non-partisan organization dedicated to conducting economic research and to disseminating research findings among academics, public policy makers, an...

nber.org

could not have written this post better myself: really appreciate it, @rajivsethi.bsky.social! As correctly pointed out, the hard question is whether diversification provides a way to moderate without censoring: the model-based counterfactuals show that partial diversification could be the answer.

Rajiv Sethi@rajivsethi.bsky.social · 2y ago

New post on the demand and supply of hate, referencing work by @ursina.bsky.social @chidiya.bsky.social @afinetheorem.bsky.social @noupside.bsky.social (cc @instrumenthull.bsky.social) open.substack.com/pub/rajivset...

Thrilled to share that my work on feed-ranking algorithms, based on a large-scale experiment I conducted with 5 million people in India, was recently featured on the Economics that really matters blog! So excited for feedback from folks in this lovely space 🦚 www.econthatmatters.com/2024/11/can-...

Can Social Media Algorithms be Regulated to Prevent Online Radicalization in Developing Countries?

We are excited to launch our Job Market Paper Series blog for 2024-2025, beginning with our very first blog post by Aarushi Kalra.  Aarushi Kalra is a Ph.D. candidate in Economics at Brown Universi…

econthatmatters.com

"By allowing people to mute content in fine-tuned and personalized ways, it reduces the incentive to mute people." A very important insight! Would depend on whether the "Discover" algorithm is content based, in which case we still end up in very segregated echo-chambers. <shameless plug for JMP>

Rajiv Sethi@rajivsethi.bsky.social · 2y ago

New post on Bluesky content moderation (cc: @noupside.bsky.social @nachristakis.bsky.social and @prossertj.bsky.social) open.substack.com/pub/rajivset...

Next up: Aarushi Kalra (aarushirita.github.io), experimenter extraordinaire! Aarushi works on development, political economy, and the digital platforms/AI. Her JMP studies a massive RCT run on one of India's largest social media platforms, which shut down the content personalization algorithm(!)

Bild

Massive experiment on India’s TikTok-like platform: Replacing personalized algorithms with randomized content reduced hate speech by 27%—but led to a 35% drop in platform usage. Toxic users adapted by sharing hate at higher rates. #econsky

Peter Hull@instrumenthull.bsky.social · 2y ago

Next up: Aarushi Kalra (aarushirita.github.io), experimenter extraordinaire! Aarushi works on development, political economy, and the digital platforms/AI. Her JMP studies a massive RCT run on one of India's largest social media platforms, which shut down the content personalization algorithm(!)