We Are Hokmah: Comparative Messenger Pilot

AI Safety has completely changed my life and I am committed to studying and learning to make AI safe for humanity. Back in May 2026, Jack Clark gave a talk in London at the Oxford University, he said "that synthetic intelligences will influence people far more than social media did." My working theory of change is built on trusted messenger networks to non technical audiences. This project is to conduct a single site pilot and test a small part of Jack Clark's claim: can an AI generated video messenger produce trust and comprehension close to a human (trusted) to a non-technical community audience. Attendees will be split into two groups: one gets material live from a human facilitator and the other group from a video that is AI generated on a shared screen. Note: A QR code will also be available for attendees to watch the video on their own devices for ease of viewing and accessibility. Both groups will take a short pre/post survey measuring what they understood and how much they trust the message and the messenger. Who explains AI risk and whether people act on it or not is important for AI Governance. Faith institutions, rural and working class communities, HBCUs, civic and social organizations are largely outside AI safety organizations, yet they their leaders are often the most trusted in communities. This pilot is a small test of whether AI generated messengers can influence behavior or action in comparison to the human. This is an exploratory test and not statistically and depending on the outcome the next step would be to test a technology generated trusted messenger at a larger scale. The output is a pre/post report of the comprehension and trust data from both groups published as a project update.