“This is an important moment for cyber defenders across industries, and we believe AI plays a critical role in hardening digital infrastructure — including systems that support elections,” the company said.
OpenAI’s five-plank approach to the 2026 midterms
OpenAI on Wednesday laid out a five-part plan aimed at reducing AI-driven election interference and assisting defenders ahead of the 2026 midterm elections. The company said the plan focuses on: spreading reliable information about voting and election results; helping with cybersecurity; watermarking deepfakes; enforcing policies that ban users from deploying its tools for election interference; and weeding out political bias in its models. OpenAI framed the package as building on earlier commitments from major tech companies in 2024 — commitments that it acknowledged some observers judged insufficient.
Tools for election officials: Codex Security and Trusted Access for Cyber
As part of its push, OpenAI said it has made two frameworks available to election officials: the Codex Security agentic framework and the Trusted Access for Cyber framework. The company also said it has been briefing both the National Association of Secretaries of State and the National Association of State Election Directors on those tools. OpenAI described these efforts as part of a broader effort to “build resilience across the infrastructure stack, including in ways that support election execution.”
Watermarking deepfakes and a new AP data partnership
OpenAI reiterated a recent technical step to label synthetic media: last week the company announced a partnership with SynthID to add watermarks to images generated with ChatGPT so they can be evaluated for authenticity. On Wednesday the company placed that capability alongside other responses and announced a new partnership with the Associated Press to share election data — a component the company identified as new in its latest announcement. Together, these moves connect content provenance (watermarks) with a news organization’s data stream (AP) and with platform-level policing.
Reactions from election security experts
Not everyone greeted the announcement with skepticism. David Becker, executive director of the Center for Election Innovation & Research, told CyberScoop that when platforms “embrace their obligation to deliver accurate information to users,” it is valuable. Becker said, “Given the prevalence and amplification of disinformation about our elections, sometimes coming from leaders in high office, it’s always a good thing when platforms and services embrace their obligation to deliver accurate information to users. It appears OpenAI is doing that with this announcement. I hope other platforms embrace this responsibility as well.”
How election officials, cybersecurity teams, and voters are affected
- Election officials: OpenAI has been briefing the National Association of Secretaries of State and the National Association of State Election Directors and is offering the Codex Security and Trusted Access for Cyber frameworks directly to officials — concrete offerings that officials can review and, if they choose, integrate into planning and defenses.
- Cybersecurity teams and cyber defenders: The company positioned its frameworks and briefings as resources to “harden digital infrastructure,” signaling that OpenAI intends to be a technical partner for teams charged with protecting election systems and the broader infrastructure that supports election execution.
- Voters and news consumers: Watermarking of images generated with ChatGPT via SynthID and a data-sharing relationship with the Associated Press are intended to make it easier to distinguish authentic reporting and election data from synthetic content — a change intended to reduce the amplification of disinformation.
OpenAI’s announcement bundles several prior commitments and technical efforts into a single, midterms-focused package and adds the AP data partnership as a newly announced element. The company framed the work as part of a wider response to warnings from government agencies, non-governmental institutes and others about AI’s potential negative impact on elections, even as those groups also promote the technology’s benefits. Whether other platforms will follow OpenAI’s lead — and whether these measures will be sufficient to curb AI-driven disinformation ahead of 2026 — remains the practical question raised by the company’s plan.
