Report: ChatGPT Shows Limited Knowledge in Sparsely Populated Areas

A report from Virginia Tech highlights “potential biases” in the ChatGPT artificial intelligence tool’s assessments of “area-specific information regarding environmental justice issues.” Ciaran Lyons explains in Cointelegraph.

According to the report, “In states with larger urban populations such as Delaware or California, fewer than 1 percent of the population lived in counties that cannot receive specific information,” whereas areas with smaller populations had lower levels of access to similar data. “In rural states such as Idaho and New Hampshire, more than 90 percent of the population lived in counties that could not receive local-specific information.” In other words, “the tool was limited in its ability to identify and provide contextualized information on local environmental justice issues.”

Lyons adds, “This follows recent news that scholars are discovering potential political biases exhibited by ChatGPT in recent times.” A press release from Virginia Tech notes that “With generative AI emerging as a new gateway tool for gaining information, the testing of potential biases in modeling outputs is an important part of improving programs such as ChatGPT.”