On Tuesday night, two AI startups, Perplexity and Elon Musk’s xAI, set out to prove their AI chatbots could handle the high-pressure task of providing real-time election updates.
The stakes were high: as millions of Americans awaited results from a close presidential race, these chatbots aimed to keep users informed with live, accurate updates. But only one of them succeeded.
xAI Stumbles Early in the Race
Elon Musk’s Grok, developed by xAI, faced challenges right from the start. The chatbot quickly made mistakes, offering misleading information about election results even before the polls closed.
This stumble highlighted a key challenge AI chatbots face when working with live events: timing and accuracy. By providing premature or incorrect information, Grok ultimately failed to build user trust. That proved that it still has a long way to go in real-time applications.
Perplexity’s Real-Time Election Hub Shines
In contrast, Perplexity’s approach was notably different. They launched an election information hub with up-to-date maps and data sourced from Democracy Works and the Associated Press.
This information was the same used by Google’s election maps, giving Perplexity’s users a reliable way to follow the race. Instead of generating potentially unreliable responses, Perplexity leaned on verified, constantly updated data, a decision that paid off.
The election hub was a bold move for Perplexity, but it worked. The AI consistently offered accurate insights and even provided historical context, helping users understand the significance of current trends.
Unlike Grok, Perplexity handled user inquiries with care, linking to trusted sources and focusing on factual reporting.
Why Did Other AI Platforms Play It Safe?
It’s interesting to note that many major AI players, like OpenAI’s ChatGPT and Google’s Gemini, avoided election coverage altogether.
This conservative approach might seem surprising, but it reflects a growing awareness of AI’s potential for “hallucinations”. (Hallucinations is a term for when AI generates incorrect or misleading information.)
For example, OpenAI recently introduced ChatGPT Search, designed to retrieve real-time information from the web. However, due to concerns about accuracy, the feature wasn’t used for election updates.
Instead, OpenAI directed users to Vote.org. That very act acknowledges that ChatGPT is still evolving and may not be ready to handle high-stakes topics like elections.
How Perplexity Prepared for the Election
Unlike its competitors, Perplexity has been testing its search and information-sharing capabilities since late 2022. This experience, combined with its partnership with verified data sources like the AP, gave the startup the confidence to tackle live election coverage.
By relying on structured data from trustworthy partners, Perplexity minimized the risk of errors. That was proof showed that AI can be a reliable resource when paired with careful planning and trusted sources.
Perplexity’s CEO, Aravind Srinivas, proudly shared that the platform saw “record traffic” on election day, a sign of how successfully the tool engaged its audience. However, Perplexity’s success also underscores a brewing tension with traditional media.
By providing election information directly in its app, Perplexity reduced the need for users to navigate to external news sites. This move potentially drew users away from media outlets’ websites.
This approach is likely to fuel ongoing debates about AI’s role in media, particularly as Perplexity becomes a stronger competitor for attention in the information-sharing space.
Leveraging Visuals to Keep Users Engaged
To enhance user engagement, Perplexity’s election hub featured interactive maps and charts, much like those seen on popular news sites. By including a color-coded electoral map of the United States and a detailed swing state tracker, these visual tools provided clear, up-to-the-minute insights on each state.
The visual approach wasn’t groundbreaking, Perplexity’s maps mirrored those on Google and major networks, but it didn’t need to be. Users simply wanted a reliable, easy-to-follow tool, and Perplexity delivered that.
Though there were some technical issues with Perplexity’s map, including occasional glitches in vote count percentages, these bugs were quickly addressed. Srinivas personally responded to user feedback on social media, demonstrating a level of responsiveness that likely further built user trust.
Handling Complex Questions with Care
When users asked about the state of the presidential race, Perplexity’s chatbot provided carefully hedged responses. These responses had been programmed to focus on verified facts. While these responses lacked the flair of a live commentator, they offered reliable, relevant information.
Perplexity was cautious with its language, aiming to inform rather than entertain. While this approach might not draw the same level of excitement as some live coverage, it made Perplexity a go-to source for people seeking straightforward information.
Despite occasional minor inaccuracies, such as referencing polling data instead of real-time votes, Perplexity maintained a high standard of accuracy. These slight missteps were quickly corrected, allowing users to follow along without significant disruption.
A Glimpse into the Future: AI and Elections
This election marked the first time AI chatbots were used as real-time sources of election information, setting a precedent for future applications of AI in civic processes. As startups like Perplexity gain experience, the hope is that these tools will continue to improve in both accuracy and reliability.
While Perplexity had deals with Democracy Works and the AP, it also pulled information from other news outlets, including CBS, CNN, and the BBC. Though the company attributed its sources, it’s unclear if these media organizations were compensated.
This raises questions about how AI platforms might navigate content partnerships moving forward, particularly if they compete directly with news organizations for user attention.
Key Differences Between Perplexity and xAI’s Election Night Strategies
Feature | Perplexity | xAI (Grok) |
Data Sources | AP, Democracy Works, various media outlets | Unknown, likely less structured |
User Experience | Election hub with maps, state tracker, historical insights | Immediate, unverified responses |
Performance | High engagement, few errors, real-time updates | Frequent errors, incorrect data before polls closed |
CEO’s Approach | Engaged with users, resolved issues swiftly | Limited communication |
Audience Impact | Reduced need for traditional media coverage | High risk, reduced credibility |
What’s Next for AI in Elections?
Perplexity’s success on election night demonstrates the potential of AI to enhance the way people access information, especially during fast-paced events like elections. Moving forward, however, accuracy and transparency will be essential for AI chatbots to gain public trust.
As companies like Perplexity refine their technology, they’ll need to balance innovation with responsibility, especially as public reliance on AI grows. Ultimately, this election may be remembered as the moment AI truly entered the arena of live news coverage.
While there’s still work to be done, Perplexity’s performance shows that AI has the potential to serve as a reliable source of information, even in high-stakes situations.