• AI companies adjusted their offerings and election guidelines ahead of Election Day.
• Perplexity has launched an election information center with information on candidates, voting logistics and results.
• Meanwhile, Google limits AI-generated election content to avoid the risks that come with it.

Election week has finally arrived. This time around, it will serve as an important test for AI companies eager to provide their users with election-related information and updates on voting results.

ChatGPT did not exist during the last presidential election in 2020. Its launched two years ago launched a wave of chatbots and generative AI tools integrated into popular consumer products like Google Search.

Before the election, the companies behind the products had to decide what restrictions, if any, they would implement on election-related AI-generated content, analyzing the associated risks.

Some, like AI startup Perplexity, are leaning into it loudly.

THE AI search engine launched a dedicated “Electoral Information Center” that uses AI to provide information on upcoming elections and results tracking. The hub includes details on voting requirements, how to find your polling location and polling times, as well as AI-summarized analysis on ballot measures and candidates, the company wrote in an ad.

“Perplexity uses a process called Retrieval-Augmented Generation to identify relevant information and summarize it in a way tailored to a user’s query,” Perplexity spokesperson Sara Platnick told Business Insider.

For election-related questions, the company uses a “curated set” of nonpartisan, verified sources, the company said. The separate hub doesn’t use knowledge stored from the model’s training data in its responses, which helps minimize hallucinations, Platnick told BI.

Perplexity will also begin offering live election updates Tuesday using data from The Associated Press. Other responses will use information from Democracy Works and other nonpartisan factual sources like Ballotpedia and news organizations, Platnick said.

Alon Yamin, co-founder and CEO of an AI-powered text analytics platform Copy leakssaid Perplexity’s approach “can certainly help filter and contextualize content,” helping users access trusted sources.

Yamin said there are potential benefits to using AI for election information. AI can provide real-time updates and help identify trends, voter sentiment and key issues.

However, Yamin said the chances of AI hallucinations and accuracy issues pose risks.

Other companies take different approaches. OpenAI will continue to integrate election-related responses into ChatGPT while providing additional functionality in light of the election.

Starting November 5, those who ask ChatGPT about election results will see a message encouraging users to consult sources like the Associated Press and Reuters, or state and local election commissions for “the most complete and up-to-date information.” , the company said in a job updated October 31.

OpenAI said it was “actively testing safeguards” put in place over the last year and monitoring any issues. It also said it would continue to adjust its safeguards, including transparency around AI-generated content and banning chatbots from impersonating real people.

ChatGPT will also direct users to the US voting information site CanIVote.org when asked procedural questions related to elections.

Anthropic’s Claude also implemented a pop-up feature that allows users who request voting information to be redirected to TurboVote, a nonpartisan resource from Democracy Works. wrote in a message.

Anthropic has other safeguards in place, including prohibiting users from using it to promote a specific candidate or issue. He said it would also prevent the chatbot from being used to generate disinformation about election laws, candidates or related topics. In its post, the company said outputs would be limited to text only to help eliminate “the risk of election-related forgeries.”

While live election updates can be helpful to users, AI-generated election content poses a number of risks.

Yamin said AI can spread biased information, misinterpret data and create false narratives. He also added that these models “are only as good as the data they are trained on,” which opens up the risk that AI-generated responses will reflect the bias of the data they are trained on.

Misunderstandings could be amplified if AI misinterprets information and presents inaccurate results, especially in a rapidly changing and closely monitored election environment, Yamin told BI.

Googlewho has experienced some High-profile AI fails with its Gemini AI and AI Overviews, has chosen to implement constraints on what users can ask its AI products about the election. Google said it’s restrict your Gemini AI chatbot to answer questions about elections “out of great caution.”

A similar search on Google’s main search product on Monday did not trigger an AI summary, showing the “Top Stories” carousel of publishers’ articles.

Google did not immediately respond to a request for comment from Business Insider.

Brad Carson, co-founder and president of the nonprofit Americans for Responsible Innovation, told BI that while Google’s approach is laudable, it is not a comprehensive solution to the problem.

Carson said combatting misinformation is “not an easy task,” and while it makes sense for companies to try to do more to limit misinformation, the government should work on legislation that requires companies AI to clearly label the information they provide to users.

“I think other products will probably fill the void that Google left, but I think it’s Google’s responsibility to try to take a step back from that,” Carson said.