Research

Fact-checker warnings work and are broadly popular, and other insights about social media moderation from an MIT expert.

Using the wrong datasets to train AI models can result in legal risks, bias, or lower-quality models. The Data Provenance Initiative’s tool can help.

Researchers studying AI, leadership, and gender equality were named to the Thinkers50 Radar list.

Once again, AI was everywhere. But research about federal spending leads the list.

Introducing “targeted friction” into AI workflows can improve overall accuracy and reduce uncritical adoption.

When enterprises address opportunities or threats immediately, they perform better. A new research briefing looks at the traits such companies share.

When software developers were given access to an AI coding tool, productivity increased — particularly among newer hires and more junior employees.

The artificial intelligence-powered “DebunkBot” reduced individuals’ belief in conspiracy theories and lessened their conspiratorial mindset, according to a new study.

When individuals assess the truthfulness of political news during an election period, their beliefs become significantly more partisan, research finds.

A new study finds a 26% decline in data storage and a 20% increase in the cost of data.