AI is becoming part of the election information ecosystem, from how audiences seek answers to how newsrooms analyze information.
But the panelists in our AP Forward session repeatedly returned to one principle: humans remain responsible for the journalism.
Arnon Mishkin, who leads the Fox News Decision Team, said Fox is experimenting with AI to identify possible patterns in polling and incoming vote data, but nothing generated by AI goes directly to audiences without human review.
Chicago Public Media is taking a similarly cautious approach. Kelly said reporting, editing and accountability remain human responsibilities, even as the newsroom explores AI for areas such as workflow efficiency and translation.
Félix raised another consideration: protecting sensitive information. Conecta Arizona limits what source material and data it puts into AI systems, particularly when working with vulnerable communities.
At OpenAI, Josh Lawson described a different side of the equation: helping users find authoritative election information while putting safeguards around political misuse. OpenAI says it will surface live AP vote counts in ChatGPT during U.S. election night and use Democracy Works for voting and registration information. OpenAI
The takeaway for newsrooms is not simply whether to use AI, but where human judgment must remain decisive.
Actionable recommendations
- Define which election workflows can use AI and which must remain human-led.
- Require human review of AI-assisted analysis before publication.
- Establish rules for sensitive source material and personal data.
Questions for leaders
- Where could AI improve our election workflow without weakening accountability?
- Are our safeguards evolving as quickly as the tools themselves?
Related articles
