Perplexity introduces an election tracker.
The artificial intelligence startup Perplexity has launched a center to monitor the results of the general elections in the United States, becoming one of the few major AI platforms to offer this feature.
Perplexity, an AI-powered search engine, has acknowledged that it can occasionally present inaccurate results or "hallucinations." Despite this, the company is looking to prove its reliability, especially regarding election results coverage. Recently, Perplexity launched an information hub dedicated to the U.S. general elections, in collaboration with data provided by The Associated Press and Democracy Works. This new resource is described as "a starting point for understanding key issues" related to the elections.
Starting Tuesday, the site will provide live updates on the elections, using data from The Associated Press to keep users informed about the presidential, Senate, and House of Representatives races, both at the state and national levels. Perplexity thanked Democracy Works for allowing access to its election API, although it has not yet clarified whether the sources are being compensated for the data provided.
Perplexity's election hub offers answers to election-related questions, such as voting requirements and polling hours, in addition to providing AI-generated summaries of ballot measures, candidates, political stances, and endorsements. While the hub is based on information from The Associated Press and the Democracy Works API, it stands out compared to other platforms that have shown reluctance to launch similar features due to fears of misinformation generated by AI.
For example, in the recent search experience of ChatGPT, OpenAI directs users to consult The Associated Press and Reuters for questions about election results. Similarly, Anthropic’s chatbot Claude and Google Gemini refrain from addressing questions about election results. Given the tendency of AI to provide erroneous information in this area, there are expectations that Perplexity's new hub will meet the accuracy standards the company claims, considering a recent study revealed that over a third of chatbot-generated responses on electoral topics contained incorrect information.
