A brand can exclude every political channel and topic from its media plan and still end up in the middle of the election.
Paid campaign ads are just one slice of the political content consumers will see this fall. The rest moves through news clips, commentary, memes, and AI-generated content. Some of it carries a label. Much of it looks like any other post in a feed.
That makes brand safety a more complex challenge in the weeks leading up to the mid-term elections.
AdImpact projects a record $11.6 billion in political media spending across the 2025-2026 cycle, an increase of 30% over the 2022 midterms. As we covered in the first blog in this series, that political spend concentrates in a short list of markets and channels. Brand safety risk doesn’t concentrate the same way.
Three things are working against your controls right now. Your exclusion list can only cover so much, platforms are moderating less than they did in previous elections, and AI generated content is outrunning the few systems built to catch it.
Your brand safety exclusion list isn’t as exclusive as you think
Meta, Google, YouTube, X, Snapchat, Reddit, Hulu, YouTube TV, Paramount+, HBO Max, and Pluto TV all accept political advertising. Apple TV+, Pinterest, TikTok, Netflix, Disney+, and Amazon Prime do not. Other platforms are unclear or have differing rules by state.
That list only tells you one thing: which platforms sell political ads. It tells you nothing about the political content already on them, or what the creators you work with are going to post.
Political ads. In places where political ads are banned and there is no user-generated content, your exposure really is low. Other platforms can be harder to plan against. On Spotify, for instance, if a podcast whose premise is “two friends talking about everything” is part of your media plan, it would not be unheard of for politics to be discussed, especially during an election.
On the platforms that do accept political advertising, you can’t “exclude” your way out of the ads themselves. Brand safety controls govern the content your ad appears next to, not the other ads in the same feed.
Political content. This is where you do have some control. Meta’s Topic Exclusion tool lets advertisers bar content tied to news and politics, social issues, crime, and tragedy. DV360 offers avoidance categories, DoubleVerify has an inflammatory politics and news category, and other platforms offer inventory tiers, keyword lists, and blocklists.
Those tools work off content categories. They classify the post, not what’s around it. You are still exposed to the comment section, the organic post below it, or the influencer whose lifestyle video takes a turn.
Creator posts. TikTok bans political ads and influencers from accepting paid political endorsements, but a influencer can still post their own point of view. A influencer who talks about your product on Tuesday might weigh in on a ballot measure Wednesday, leaving your posts a couple of swipes away and joined by association.
And a influencer can take political money without ever telling you. That’s its own problem, and we’ve written it up here.
Platform moderation has relaxed and your brand safety controls cover less
Since the last election cycle, platforms have rolled back the rules that governed political content.
YouTube retired its 2020 election integrity policy in 2023. In September 2025, Alphabet told the House Judiciary Committee that YouTube’s guidelines now allow a wider range of content on elections integrity, and that creators banned under the retired policies could apply to come back. Reinstatements started in October of that year.
Meta ended its US third-party fact-checking program in 2025, moving lower-severity enforcement to user reports while increasingly recommending political content to people who show interest.
The Center for Democracy and Technology’s midterm analysis found that changes at both platforms leave these elections with greater latitude for hateful and violent content than the last cycle.
Your filters only remove what the platform has already flagged as a violation. Platforms are flagging less now and most flagging is now done through user notes, rather than from the platform itself. Your settings haven’t changed, but they’re catching less than they did in 2024.
AI raises the speed and the stakes of brand safety
AI-generated content costs almost nothing to make and spreads faster than anyone can catalog it. Your blocklists, keyword filters, and verification vendors all work by matching a list of things already identified. A fake that went up this morning will not be part of that list.
NewsGuard found 393 programmatic ads from 141 major brands on AI-written news sites over two months in 2023. More than 90% ran through Google. NewsGuard didn’t name the brands, but the list covered banks, luxury department stores, and streaming services.
Picture your YouTube buy landing next to what looks like local news coverage of a Senate race, except it’s AI-generated, with image and voice deepfakes. The platform hasn’t caught it yet, and your spot keeps running.
That placement costs you. In an experiment by Zefr, with 2,000 US participants, 47% said adjacency to misinformation compromises a brand’s integrity. Only 36% of participants could correctly identify AI-generated misinformation.
Imagine you have a multi-year deal with a well-known actress. In October, her voice shows up in a political attack ad. She never recorded a word of it. She can sue under right of publicity laws, in some states, but that doesn’t help you, because your campaign featuring her is running the same week.
Thirty-three states now regulate deepfakes in political campaigns. Most require a label. A few ban deceptive AI outright, and those have fared worst in court. Judges blocked the bans in California and Hawaii for failing to exempt satire and parody. Only a few states let people sue over their face or voice. It’s up to each campaign to work out which rules apply.
AI disclosure rules skip the channels with the biggest political spend
Google and Meta both require political advertisers to disclose AI, which both platforms did voluntarily. Google makes verified election advertisers disclose synthetic content that depicts real or realistic-looking people. Meta requires the same for ads about social issues, elections, and politics. Since June 2026, Meta has been running automated AI detection software.
Broadcast television, the primary driver of election spend, has no rule at all. According to AdImpact, broadcast is expected to receive 48% of political spend this cycle. It’s the one channel with a federal regulator, and it’s the one channel with no AI disclosure requirement.
The FCC proposed an AI disclosure requirement in July 2024, which would have required TV and radio stations to ask whether a political ad used AI and disclose it on air. It advanced over dissents arguing the agency lacked authority. One of those dissenters, Brendan Carr, became FCC chairman six months later. To date, the FCC has not adopted a final rule.
State law doesn’t close the AI disclosure gap either. California’s AB 2355 specifically exempts radio and TV stations when they’re paid to air a qualified political ad, even if that ad is AI-generated. The same spot online would require disclosure and might even be banned.
CTV, the fastest-growing political medium in the country, has no regulatory or central platform policy.
What to do about brand safety this election
Plan for content your filters can’t identify. Consider more intentional media buys, including private marketplaces or publisher direct placements, where there are higher levels of curation.
Your account defaults might be built for a different election year, so audit your settings at a more granular level. And identify the person who can pause a campaign before you need them.
By being proactive around the topics and situations that matter to your brand, you can act before a bad situation makes the decision for you and you are reacting from the back foot.
Our job is to be a neutral steward of your media investment. That means an unvarnished read on where the risk actually sits and what each option costs you. Where and how your brand shows up during an election is your decision, and it should be.
Want to pressure-test your brand safety plan before the midterms peak? Contact us.
This is the second blog in our midterm election series. Follow us on social media more exclusive insights.
