Generative AI is changing how people look for election information.
A voter can now ask ChatGPT, Gemini, Claude or Grok where to vote, how ballots are counted, whether an election technology is secure, what happens during a technical issue, or whether a fraud-related rumor is true.
The answer may be based on official information. But it no longer comes directly from the election authority. It comes through a system that selects, summarizes, and reframes information from multiple sources.
That creates a new challenge for election authorities: what is AI telling people about elections, election authorities and the safeguards that protect electoral integrity?
And just as important: what sources is it using?
In elections, the source matters as much as the answer
The risk is not limited to an AI model providing incorrect information. A response can be technically accurate but incomplete. It can omit jurisdictional context, mix procedures from different countries, or rely on secondary sources even when official information is available.
In elections, those nuances matter.
An ambiguous explanation about vote counting can fuel suspicion. An incomplete answer about where to vote can affect participation. A statement without context about audits, results transmission or election technology can weaken public trust.
That is why election authorities need to expand what they monitor.
It is no longer enough to track media coverage, social media conversations and disinformation campaigns. Election authorities also need to understand how AI models respond to citizens’ real questions, and whether official sources are central to those responses.
When the official source is missing
When institutional information is clear, current, verifiable and easy to find, AI models are more likely to use it as a reference.
But when election authorities do not publish enough information, when their content is difficult to find, or when it does not answer the questions citizens are actually asking, that space can be filled by media outlets, blogs, political actors, forums, social media posts or outdated content.
In that scenario, the election authority does not only lose visibility. It also loses the ability to shape the institutional framework through which its processes, safeguards and results are understood.
The challenge, then, is not only to fight disinformation. It is also to avoid the absence of official information in the spaces where citizens are already looking for answers.
Building presence before the crisis
This was one of the ideas presented by Samira Saba, Smartmatic’s Director of Communications, during the International Capacity-Building Program on Electoral Management: Innovation, Integrity and Democratic Governance, held at the University of Salamanca on June 29 and 30, 2026.
Her remarks were part of the panel “Disinformation and the Role of Election Management Bodies,” organized by Transparencia Electoral in partnership with the University of Salamanca’s DEGAM Knowledge Transfer Group.
One of the key takeaways is that institutional presence cannot be improvised during a crisis. It must be built in advance through official information that is clear, accessible, verifiable, up to date and designed to answer the questions citizens are actually asking.
In the age of generative AI, publishing accurate information remains essential. But it is no longer enough.
The new challenge is to become the source AI chooses to use.