On Tuesday, two AI startups tried to convince the world that their AI chatbots were good enough to be an right source of real-time information during the high-stakes presidential election: xAI and Perplexity.
Elon Musk’s grok failed almost immediately, giving incorrect answers about the race results before polls even closed.
Perplexity, on the other hand, provided useful real-time election information and maps throughout the night, connecting to credible resources and offering historical context where appropriate.
Embarrassed, I decided to take on a risky venture and it paid off.
Slow last week, the startup announced the launch of an election information hub with real-time maps featuring voting data from Democracy Works and the Associated Press, the same information sources that power Google’s election map. This approach was different from most other AI chatbots, such as OpenAI’s ChatGPT or Google’s Gemini, which simply refused to answer election-related questions.
It’s understandable why most AI labs waited out this election. For many of them, this was a unthreatening and responsible choice because they had been plagued by embarrassing hallucinations at some point over the past year.
Notably, OpenAI recently released its Google competitor, ChatGPT Search. However, the startup led by Sam Altman did not trust this feature to answer questions about this election, directing users to Vote.org instead. The ChatGPT search engine is still an early product that is too unreliable for people to operate in everyday situations, and OpenAI seems to have noticed this.
Perplexity, on the other hand, has been testing its Google competitor in the real world since December 2022 and clearly felt it had enough data to give these choices a chance.
Perplexity Club’s success on election night could hamper its ongoing fight against media companies; most notably the recent Dow Jones lawsuit, which alleged that the startup competes with media companies for the same audiences. Despite the many outbound links in the AI chatbot’s responses, Perplexity’s voting center itself was an election night destination and did not require users to leave the app to get all the information. It certainly appears that Perplexity was competing with media companies that were also vying for eyeballs on election night, even though Perplexity collects information from these media outlets. Even CEO Aravind Srinivas is embarrassed he claimed the day before, “record traffic” and clearly hoped to maintain this momentum.
While Perplexity worked with Democracy Works, AP and several other media companies to power election materials, the startup also indiscriminately used live election coverage from other media outlets such as CBS, CNN and the BBC. Sure, Perplexity offered attribution, but the company hasn’t announced any revenue-sharing partnerships with these outlets, and it’s unclear if any money changed hands.
How embarrassment dealt with election night
First, let’s start with Perplexity’s election features that have nothing to do with generative AI: graphs.
People usually love visual election charts, they click on them and see detailed state-level data. Creating them was a shrewd move on Perplexity’s part, ensuring that their AI systems weren’t the only source of information in the app.
Visiting Perplexity’s voting center, users saw a familiar-looking US electoral map, with some states blue for Kamala Harris and others red for Donald Trump. Of course, Perplexity didn’t reinvent the wheel with this feature – copying Google’s display and every TV network’s programming – but there was no need to. Throughout the night, that map appeared to update about every minute, according to information posted on the Associated Press website. It was a good way to follow the elections.
Throughout the night, the Perplexity map experienced periodic errors. Sriniwy replied to users with X that reported bugs such as Perplexity missing information about the percentage of votes counted and quickly fixed them.
Perplexity offered another famed feature, a state-by-state tracker, providing real-time information about swing states.
Now the artificial intelligence part. Perplexity responded to questions about the current state of the presidential race with hedged answers that still provided mostly right information. These answers weren’t as insightful as a CNN commentator, nor as entertaining as The Novel York Times’ election needle (which returned this year). However, Perplexity only showed a few diminutive hallucinations and largely presented the relevant facts in a timely manner. That’s more than I can say for any other AI chatbot on the market.
While trying to answer further questions about Harris’ role in the Blue Wall states, Perplexity had delicate hallucinations. He was referring to polling data when he should have been referring to real-time votes at that time of night. However, the general information here was right and other AI chatbots simply wouldn’t answer this question.
Here’s another follow-up question we asked: What ballots haven’t been counted yet in swing states? It was complex to find the answer to this question elsewhere. True, the only useful answers were for Pennsylvania and North Carolina, but at least for the remaining states, Wonder wasn’t a hallucination.
This election marks the first time that AI chatbots have been used as a source of information about our democratic processes. However, these will not be the last elections in which such a situation will take place. Well-funded artificial intelligence startups are fighting to deliver information to people in modern, faster and more concise ways. Accuracy will be the key to success in the future. For now, Perplexity has the early advantage.