Using AI in advocacy communications
A couple of quick experiments- one is on misinformation, the other on immigration
I wanted to play around with ChatGPT for a couple of experiments. Let me say at the top: If you’re already well-versed on AI, this is rudimentary. It’s for beginners who are interested in how to apply the technology.
You may have heard how important the prompt is. This is a general starter piece on prompts from MIT and this story gives specific examples of prompts for comms work. There’s a whole strategy to it, which I’m still learning about.
Experiment 1: Ad script
I asked ChatGPT to make a one-minute web ad script. Here’s the prompt I provided:
Create a one-minute web video on the benefits of using AI to find misinformation and disinformation in a social media feed. I want the name of the ad to be “Lie Detector” and use that metaphor throughout the video.
Here’s verbatim what it generated.
🎥 VIDEO TITLE: "Lie Detector: AI vs the Firehose of Falsehood"
[0:00–0:05] — Cold Open / Hook
Visual: Rapid flashes of fake headlines, AI-generated deepfakes, tweets spreading panic
NARRATOR (VO):
"Your feed is a war zone of lies. But what if you had a Lie Detector in your pocket?"
[0:06–0:15] — Introduce the Metaphor
Visual: Retro-style polygraph morphs into sleek AI interface scanning a social feed
NARRATOR (VO):
"AI can scan thousands of posts a second—flagging false claims, manipulated images, and shady sources before they go viral."
[0:16–0:30] — Real-World Use Cases
Visual:
– Heat map of disinformation spreading
– Side-by-side: flagged vs. unflagged content
– Dashboard showing real-time threat detection
NARRATOR (VO):
"From foreign propaganda to fake medical cures, AI-powered lie detection is helping journalists, fact-checkers, and watchdogs fight back."
[0:31–0:45] — Emotional Core / Stakes
Visual: Teen scrolling through TikTok, a parent reading false vaccine info, a voter misled before election day
NARRATOR (VO):
"Lies online don’t just mislead—they divide us, hurt people, and even cost lives. Truth needs backup."
[0:46–0:55] — Call to Action
Visual: A “Lie Detector” icon appears on a screen. Fact-checkers and AI work side-by-side.
NARRATOR (VO):
"AI is your new truth ally. Not to replace your judgment—but to sharpen it."
[0:56–1:00] — Outro
TEXT ON SCREEN:
“Lie Detector: Because the truth deserves tech on its side.”
[Optional logo/URL if this is part of a campaign]
I like the first two-thirds. I don’t love the “call to action” language or the “outro.” Too many people already don’t trust or are wary of tech. And that last line is just cheesy. I might end the spot instead with something like “fight back against the misinformation makers/don’t get fooled/ turn the tables on them.” Makes it more empowering.
Experiment 2: Public sentiment analysis
AI is praised for its public sentiment analysis. It looks at content trends to see, in my words, which side is winning the messaging war. So on June 14 I queried ChatGPT again to see which side of the immigration debate is performing better on social media, excluding X. That platform is so anti-immigrant it would skew the results.
Here what it generated:
The anti-immigration side has the upper hand on social media—dominating volume, speed, and influence—though pro-immigrant voices gain traction during key flashpoints.
This tracks what’s happened in the real world over time. Democrats have not made immigration reform a top priority over many years and, unlike the Democratic Party broadly, the issue has received disproportionate attention on the right. While the issue of immigration is nuanced, polling and the 2024 election outcome show conservatives tend to have an advantage.
Here is a fleshed-out version from ChatGPT:
Fear-driven narratives gain early lead: Pro-Trump influencers, as well as government and extremist voices, have successfully spread sensational anti-immigration content—framing migrants as threats or invaders—which consistently goes viral, especially on Instagram and Facebook.
Official meme warfare: The Trump administration, DHS, and conservative figures actively use memes celebrating deportations—and these receive notably high engagement across Instagram and YouTube.
Pro‑Immigrant Pushback: Strong presence during raids and protests: Emotional videos and images showing families affected by ICE raids, or solidarity rallies (including cultural pride expressions in L.A.), often go viral on YouTube, TikTok, and Instagram.
Grassroots mobilization: Migrant communities and allies use Facebook and TikTok as real-time alert systems (#PeopleOverPapers), bolstering pro-immigrant reach—though reactionary compared to anti-immigrant messaging.
My next experiment is creating an AI presidential candidate from whole cloth, giving her the attributes, policy positions, and strategy to win a nationwide election. Stay tuned.