AI Briefing: How political startups are helping small political campaigns scale content and ads with AI

With about 100 days until Election Day, politically focused AI startups see the flourishing technology as a way to help national and local candidates quickly react to unexpected change — while also putting the right guardrails in place. 

Since privately launching this past spring, BattlegroundAI has helped state- and local-level candidates create YouTube scripts, social content and digital ads using generative AI. Powered by top large language models — including Gemini, ChatGPT, and Claude — BattlegroundAI’s platform is helping solo campaigns and small organizations benefit from top AI models to craft content and analyze data quickly and at scale.

After working in both advertising and in politics — including former President Barack Obama’s 2012 campaign — BattlegroundAI Founder and CEO Maya Hutchinson wanted to teach political campaigns how to use AI in advertising while also building a platform to serve as a “central nervous system” for streamlining adoption of AI-powered ads. Next month, the startup is getting ready to launch a public beta program that will be priced based on scale and volume of content, starting with five messages for free or a subscription with unlimited messages for $19 a month.

“This is like jump-starting that process, especially as you can see right now, there’s going to be a huge influx in fundraising and money coming into these races,” Hutchinson told Digiday. “A lot of that will go to advertising, so how can you quickly spin up 100 search ads?…The volume is going to be high and the variations are going to be really important.”

Battleground is just one of several AI startups to have sprung up in the past year. Another company Quiller, founded a year ago this month, began with features for AI-generated email campaigns but has since then expanded to helping campaigns with AI-generated copy for social media, digital ads, press releases and op-eds. Meanwhile, others are using visual AI and LLMs to generate AI to create audience segments for CTV political ads and other online platforms.

According to Quiller co-founder Mike Nellis, AI “is revolutionizing political campaigns, but in the most boring way.” Using AI to write emails faster helps with fundraising but also with drafts responding to major events either related or unrelated to a candidate. “The first person to get that tweet out, the first person to get that fundraising email out, the first person to get that press release out — they are usually the person who gets the most attention.”

Small campaigns that are often “armies of one” aren’t necessarily able to afford spending a lot to hire content and media agencies contracted by national and state-level campaigns. One example Nellis gave is a mayoral candidate in Bowling Green, Ky. — who is running a campaign with just one volunteer — who is using Quiller to cut down the time it took to write fundraising emails. However, Nellis said small campaigns are also more likely to adopt new tech than national and state-level efforts. That’s partially because many of the headlines related to AI-generated political ads have been about “worst possible uses cases of AI” like deep fakes.

“Both political parties are terrified to use it and [are] being very tepid about it, and they’re not leaning into innovation in the way that they should be,” added Nellis.

Not all agencies working with political candidates are using AI yet for content. Instead of focusing on AI for content-generation, Tyler Goldberg, director of political strategy at Assembly Global, thinks the biggest growth market for AI is for building contextual audiences. While he hasn’t used AI for writing emails yet, he sees how it might be helpful to reach new audiences and bring more voters into the democratic process. 

Just like past tech innovations in politics, Goldberg thinks AI will soon become just another tool that also comes with new responsibilities around privacy. Although regulation and transparency are needed for AI, he said a blanket ban on using the tech in politics is just “low hanging fruit” that misses some of the benefits. 

“I’ve volunteered on campaigns before, used to physically dial phone numbers and even had a headset — and that wasn’t all that long ago,” Goldberg said. “…“There was a time when having a political ad on TV was thought to be a wild thing that would never happen. Same with radio, same with buttons I’m sure at one point or another.”

More companies in the ad tech space are already using AI to create new audiences based on consumer and voter data. Today, Yahoo and Resonate announced plans to integrate Resonate’s voter segments with Yahoo’s demand side platform, giving advertisers access to more than 1,000 audience segments based on political affiliation, interests, shopping behaviors and health categories.

“Even for national campaigns, because of how the voting season is designed, most of the budgets are still getting allocated for very small geographies where the race is closer to 50-50,” said Giovanni Gardelli, Yahoo’s vp of ads data products. “And both parties are trying to pick the voters that are undecided yet in the middle. Regardless of the level at which selections are being run, we see a majority of budgets being very targeted to smaller regions.”

The news comes just days after Resonate released its annual list of 10 voter segments for the 2024 election that shed light on the types of audiences campaigns might look to target based on political learnings and various data sources. “Progressive Professionals” and “Safety-Net Seniors” are both left-leaning groups focused on social and environmental issues, with the former active on social media and the latter preferring traditional media. “Non-Voting Neutrals” are politically apathetic Gen Xers, unregistered to vote, financially pressured, with limited media engagement. “Money-Minded Undecideds” are retired Baby Boomers earning under $25,000 annually, focused on saving money, healthcare, gun control, affordable medications, and corporate oversight.

The bowing out of incumbent U.S. President Joe Biden from the race is “such a dramatically different situation” that could lead political advertisers to reach voters they might not normally target, said Resonate CMO Erika McCoy. While many voters and donors have been “sitting on the sidelines with their money,” she said that could change fast with more money to be raised — and more money to be spent. 

“Everyone has really held off thinking this isn’t a race they need to invest in, but I think we’re going to see now that it is,” McCoy said. “They are going to need to get certain segments of voters to move one way or another… There are more people now up for grabs that can be persuaded based on various interests and groups. Those are going to be the people [campaigns] have to get moving quickly.”

Prompts and Products: AI news and announcements: 

  • OpenAI is testing an AI-powered search engine prototype called SearchGPT that could rival giants like Google.
  • Publicis Groupe is getting deeper into the creator economy by acquiring Influential to integrate the startup with consumer data from Epsilon. The plan is to use AI across both platforms to offer advertisers an enhanced creator network, ID-based content and cross-channel media to better match creator content with right audiences.
  • Meta announced its latest AI model, Llama 3.1, and debuted new features for Meta AI including seven new languages, new creative tools that include editing tools and a way to make personalized AI-generated images, as well as expanded access to Meta AI via Meta Quest.
  • The FTC is looking into whether eight companies are offering “surveillance pricing” to set prices based on personal data: location, demographics, credit history, browsing/shopping history, etc. Companies that received letters were Mastercard, JPMorgan Chase, Accenture, McKinsey & Co., Revionics, Bloomreach, Task Software & PROS.
  • Condé Nast has reportedly sent Perplexity AI a cease-and-desist order over allegations of data-scraping without permission.
  • Amazon announced new tools for AWS to help contact service agents use generative AI to field interactions and analyze data.
  • The AI model provider Cohere raised another $500 million for its enterprise focused AI models while also reportedly cutting some staff.
  • Brandtech Group named James Dow to the newly created role of Gen AI creative director.
  • Officials from the U.S., European Union and U.K. issued a joint statement to look into antitrust allegations related to AI competition.
  • Former NFL star Colin Kaepernick announced a new AI startup called Lumi Story AI to develop a creator-focused platform for generating text and visuals. It’s being backed by Seven Seven Six, a VC firm founded by Reddit co-founder Alexis Ohanian.

1s and 0s: AI-related research and reports

  • Gartner published its Digital Marketing 2024 hype cycle report, which details continued adoption of generative AI along with the use of “emotion AI” to engage people with chatbots using natural language processing.
  • Google’s quarterly earnings were once again dominated by dozens of mentions of AI. However, its news of not deprecating cookies only came up once.
  • Canva and Harvard Business Review released a new data report about creativity and using AI at work.

Other Digiday stories related to AI:

https://digiday.com/?p=551043

More in Media

Incoming teen social media ban in Australia puts focus on creator impact and targeting practices

The restriction goes into effect in 2025, but some see it as potentially setting a precedent for similar legislation in other countries.

AI Briefing: Amazon’s new Nova models boost AI model efficiency, accuracy and variety across AWS

One of the most buzzy debuts was Nova, a suite of six new AI models that include understanding and creating text, images and videos.

Q&A with Jessica Chan, Perplexity’s head of publisher partnerships

Perplexity’s new head of publisher partnerships Jessica Chan shares how the AI tech company is wooing publishers, from what the program offers now to what she hopes to add to the program next year.