As Americans turn out to the polls on Tuesday to vote, AI startup Perplexity is launching an election information hub, powered by data from the Associated Press, to provide live updates on the results.
“We want to do our part to support an informed electorate,” the company announced Friday, “so we’ve built the election hub on Perplexity’s answer engine: an entry point for understanding key issues, voting intelligently, and tracking election results.”
The hub will cover the U.S. presidential election and senate and house contests at both the state and national level. You’ll be able to ask about a variety of election-related topics such as “voting requirements, how to find your polling place, and poll times,” per the company, “as well as receive detailed, AI-summarized analysis on ballot measures and candidates, including official policy stances and endorsements.”
Perplexity is relying on “a curated set of the most trustworthy and informative sources,” specifically from the AP and Democracy Works (which also powers Google’s election search results). “We selected domains that are non-partisan and fact-checked, including Ballotpedia and news organizations,” a Perplexity representative told The Verge. “We’re actively monitoring our systems to ensure that we continue to prioritize these sources when answering election-related queries.”
Perplexity’s election efforts are unique among its chatbot peers. OpenAI’s ChatGPT will not answer election-related queries, instead directing users toward the AP and Reuters for information. Both Google’s Gemini and Anthropic’s Claude similarly refuse to engage with questions on that topic, likely in order to avoid the potentially embarrassing and damaging “hallucinated” responses that continue to plague today’s generative AI models.
A July study from the Center for Democracy and Technology, for example, found that more than a third of responses to 77 election-related questions asked of AI chatbots contained incorrect details and misinformation.
Perplexity itself is currently being sued by News Corp over allegations of “massive copyright violations” and that the company’s chatbot had scraped news, opinions, and analysis directly from the Wall Street Journal and New York Post, both of which News Corp owns.
The publisher also claims that Perplexity’s AI routinely hallucinated facts in its replies, “sometimes citing an incorrect source, and other times simply inventing and attributing to Plaintiffs fabricated news stories.”
Perplexity has also been accused of “willful infringement” of Forbes‘ content and has been sent cease and desist letters from both The New York Times and Conde Nast over its sticky-fingered behavior.
Read the full article here
