The 2024 election cycle is the first AI deployed by a political movement. Although candidates largely avoid major unfortunate events, the technology has little guidance or limitations. Now, the National Democratic Training Council (NDTC) is launching its first official script, where Democratic campaigns can use AI responsibly before the midterm.
In a new online training, the committee has developed a plan to enable Democratic candidates to use artificial intelligence to create social content, write voter outreach information, and study their regions and opponents. The group said it has trained more than 120,000 Democrats to seek political positions since its inception. The group will become democratic politicians with virtual courses and in-person training, from voting registration and fundraising to data management and on-site organization. The group is largely targeting smaller events, with less resources in its AI courses, trying to empower a five-person team that works with “Efficiency of a 15-person team.”
“AI and AI adoption of the person in charge is a necessity for competition. It’s not a luxury,” said Donald Riddle, senior teaching designer at NDTC. “It’s something we need learners to understand and feel comfortable implementing so that they have a competitive advantage, drive incremental change, and push the left and right push to the needle while using these tools effectively and responsibly.”
The three-part training includes explanations about how AI works, but the course meat revolves around possible AI use cases. Specifically, it encourages candidates to use AI to prepare text for texts for a variety of platforms and uses, including social media, email, speeches, telephone banking scripts and in-house training materials that are reviewed by humans before publication.
The training also points to ways that Democrats should not use AI and prevent candidates from using AI to slam their opponents, imitate real people or create images and videos that could “deceive voters through false statements, individuals or reality).
“This undermines the discourse of democracy and the trust of voters,” the training reads.
It also advises candidates not to replace human artists and graphic designers with AI to “maintain creative integrity” and support work creativity.
The final part of the course also encourages candidates to disclose AI usage when content has AI-generated voices, either for “deep personalization” or for crafting complex policy positions. “When AI significantly promotes policy making, transparency builds trust,” it wrote.
These disclosures are the most important part of Hany Farid, an AI expert and professor of electrical engineering at the University of California, Berkeley.
“When something isn’t real or completely generated, you need to have transparency,” Farid said. “But the reason for this is not only because we reveal something that is not real, but also what we believe in is real.”
When using AI for videos, NDTC recommends campaigns to use tools like descriptions or work editing to script and quickly edit content for social media, stripping long videotapes and awkward moments.
