AI content moderation takes a lesson from economics #ASA190
Reports and Proceedings
Updates every hour. Last Updated: 13-Jun-2026 14:15 ET (13-Jun-2026 18:15 GMT/UTC)
Large language models may be able to make moderation of social media and online communities more effective, but they are expensive to run at scale, especially when asked to provide explanations for each piece of content they flag. Yuan Zhao will present research on creating an interpretable and low-cost method for evaluating LLMs’ hate speech classification as part of 190th ASA Meeting. His framework relies on the Rational Inattention model, an economic idea developed to explain human behavior.
A new UC Irvine study uses monetized life-cycle analysis to compare hydrogen, direct electrification and fossil fuel pathways across heavy-duty transport and industrial sectors. The researchers concluded that renewable hydrogen applied in certain sectors offers greater social value, which includes reduced climate change impacts, cleaner air, improved public health and lower demand for natural resources. The researchers pinpointed steel, transoceanic shipping and long-haul trucking as the highest-value targets for clean hydrogen deployment.
In most U.S. zip codes, electric vehicles are cost-competitive with their gas-powered counterparts, according to a new study. And regarding emissions benefits of EVs, individual driving patterns matter as much as regional factors like the local electricity mix.