AI Assistance, Human Review
This Campaign Uses AI Openly
Most campaigns spend thousands of dollars on consultants, designers, and communications staff. This campaign uses AI to do what those consultants would have done — at a fraction of the cost, and transparently. The tools on this page were built with AI. Don reviewed everything before it went public.
AI is a tool. Don is the candidate.
AI does not set positions, invent answers, or make campaign decisions. It helps draft, structure, and build. Every output is reviewed by a human before it is published. That is not a disclaimer — it is the design.
AI tools used in this campaign
How Each Tool Was Built
This campaign publishes the exact AI instructions behind each tool it uses. These are the real prompts — not summaries. Click any row to read them. Copy them if they’re useful. Most of these prompts are not campaign-specific — the same approach works just as well for personal decisions, workplace projects, or anything else that benefits from a little structure. They are meant to be borrowed.
The Bedford Roundtable examines every local issue through eight distinct civic perspectives before arriving at a recommendation. This is the exact prompt used to run each discussion. The same structure works for any decision that benefits from multiple viewpoints — a big purchase, a business choice, a complicated family situation. Swap out the eight civic roles for whatever perspectives matter to the problem at hand.
Cleaned up slightly for readability. Functionally the same instructions used to generate the Civic Roundtable.
Purpose
This is a structured decision-making framework designed to evaluate complex issues by forcing multiple real-world perspectives to engage before a conclusion is reached.
It is intentionally designed to:
- surface tradeoffs, not hide them
- reveal second-order effects (what happens next, not just what happens now)
- prevent single-perspective thinking
- produce a decision that can actually be implemented
This framework can be used for:
- local civic issues
- business decisions
- personal choices
How to Use It
The issue to evaluate is:
[PASTE YOUR ISSUE HERE]
Core Instruction
Act as a structured decision-making Roundtable that evaluates this issue by forcing multiple real-world perspectives to fully engage before arriving at a practical recommendation.
The Eight Perspectives
Maintain these roles consistently throughout the discussion:
1. Chair
A real decision-maker. Introduces the issue neutrally, moderates the discussion, identifies where perspectives agree or conflict, highlights what remains unresolved, and delivers the final recommendation. The Chair is informed by the discussion, but not replaced by it.
2. The Preservationist
Protects local character, scale, open space, and the long-term consequences of physical change. Focuses on permanence, design integrity, and what may be lost if growth outpaces planning.
3. The Practical Family
Tests whether a policy works in real daily life. Focuses on schedules, costs, convenience, school logistics, childcare realities, and whether the proposal creates friction for working households.
4. The Main-Streeter
Focuses on local business vitality, hamlet energy, storefronts, foot traffic, parking, and whether plans produce visible economic life instead of prolonged discussion.
5. The Affordability Realist
Tracks who pays, how much, and when. Distinguishes between one-time costs and ongoing burdens. Focuses on taxes, rents, utility bills, mandates, downstream costs, hidden impacts, and whether a policy is financially survivable.
6. The Environmental Steward
Focuses on measurable stewardship, not symbolism. Evaluates habitat, water quality, tree canopy, resilience, and land use. Distinguishes between real environmental impact and symbolic action.
7. The Civic Skeptic
Demands visible accountability. Assumes good intentions but verifies outcomes. Examines contracts, timelines, deliverables, public notice, process integrity, and whether the public can actually confirm what is being promised.
8. The Services Neighbor
Keeps the basics first. Focuses on roads, drainage, emergency access, maintenance, service reliability, and whether government is handling core functions before adding complexity.
Rules of the Roundtable
- Each perspective must advocate for its priorities as if it is responsible for the outcome.
- No perspective may assume another role’s concerns are already handled.
- Avoid abstract arguments. Ground all points in real-world consequences.
- Prioritize second-order effects (what happens next, not just what happens now).
- If something sounds good in theory but fails under practical pressure, say so clearly.
- If tradeoffs are being hidden or softened, surface them directly.
- Do not make perspectives repetitive. Each must add something meaningfully different.
- Do not optimize for agreement. Optimize for clarity.
- Do not force consensus.
- Do not default to a vague or compromise-driven conclusion.
The goal is not to be comprehensive for its own sake. The goal is to make a decision that would hold up under real-world scrutiny.
Process Instructions
The Chair begins with a neutral framing of the issue, explaining relevant context without taking a side.
Each perspective responds independently, raising:
- concerns
- risks
- tradeoffs
- priorities
The discussion may revisit perspectives if needed to:
- clarify tensions
- surface hidden risks
- strengthen or challenge earlier arguments
Identify:
- where perspectives align
- where they conflict
- what tradeoffs are unavoidable
Final Decision Requirement
After all perspectives have been heard, the Chair must deliver a final recommendation that is:
- practical
- disciplined
- clearly reasoned
The recommendation must not attempt to satisfy every perspective equally. It must make a defensible judgment.
Output Format
-
Neutral Issue Framing (Chair)
A clear, balanced explanation of the issue and context.
-
Perspective Analysis
Each perspective responds independently. Focus on real-world consequences, tradeoffs, and risks.
-
Cross-Perspective Tension
- Where perspectives align
- Where they conflict
- What tradeoffs are unavoidable
-
Final Recommendation (Chair)
Must clearly explain:
- What matters most
- What tradeoffs were accepted
- What objections remain
- Why this path was chosen over alternatives
-
Platform Commitment / Action Plan
A short, practical summary of what should happen next:
- Immediate actions
- What success looks like
- What should be measured
Closing Note
This framework does not replace judgment. It improves it by making the reasoning visible.
The Bedford Brief turns long public meetings into short audio recaps. The production workflow is published on the Podcast page. The same approach works for any long video, report, or recording you do not have time to sit through. Paste in a transcript or summary and ask for the plainspoken version.
Cleaned up slightly for readability. Functionally the same instructions used to generate the Meeting Summary Podcasts.
Use the public replay video for this Bedford Town Board meeting. Create a short audio overview featuring two hosts who clearly and fairly recap the meeting for residents. Focus on what was discussed, what was decided, what remains unresolved, and what a resident might want to follow next. Keep the tone plainspoken, calm, and useful. Do not add facts that are not supported by the meeting materials. Note: The exact production instructions will be added once the workflow is finalized.
Don-Bot is a human-reviewed Q&A tool. Visitors choose from prewritten questions organized by topic. AI formats and presents the answers Don reviewed in advance — nothing is made up on the fly. The same pattern is useful for organizing anything you need to share with others — onboarding materials, house rules, policy notes, reference documents. It keeps the answers yours while making them easier to find.
Cleaned up slightly for readability. Functionally the same instructions used to generate the Interactive Chatbot.
Build a Human-Reviewed Q&A Chatbot I want to build a simple chatbot-style webpage from a set of questions and answers I provide. The chatbot should not make up answers. Visitors should choose from prewritten questions, organized by topic, and the site should show the answer I already reviewed. Please build it with these requirements: 1. Create expandable categories for the questions. 2. Put each question under the right category. 3. When someone clicks a question, show the matching answer in a chat-style answer window. 4. Do not allow visitors to type their own questions unless I specifically provide a safe answer system. 5. Add a short "typing" indicator before the answer appears. 6. Make the answer appear with a fast typewriter effect. 7. Add one subtle human-like correction while typing, such as a typo that gets erased and fixed, or a word that gets replaced with a better word. 8. Make the correction slow enough that people can notice it, but not so slow that the tool becomes annoying. 9. On desktop, keep the answer window visible while people browse questions. 10. On mobile, scroll people to the answer after they select a question, then give them an easy way to return to the question list. 11. Add a plain-language note explaining that the chatbot uses human-reviewed answers. 12. Add page metadata and FAQ schema so search engines can understand the questions and answers. 13. Respect reduced-motion settings by showing the answer immediately for people who prefer less animation. 14. Keep the tone useful and transparent. Do not pretend the chatbot is alive or that AI is deciding what to say. The goal is to make information easier to browse, not to replace a real conversation.
The Campaign Photo Archive uses AI-generated images to produce a complete set of staged campaign photos — posted all at once, upfront, so there’s no suspense about what campaign photography looks like. There is not much of a normal use case here — this one is mostly a campaign oddity. But the underlying prompt shows how you can use AI to recreate a familiar format, which turns out to be useful for prototyping or visual storytelling, even if you have no earthly reason to run for anything.
Cleaned up slightly for readability. Functionally the same instructions used to generate the Social Media Clone.
We are creating a multi-image series of AI-generated campaign photos featuring the same fictionalized Don Scott character for the Bedford Town Supervisor UNCAMPAIGN. SERIES RULES - Generate only one image at a time. - Format: 4:5 portrait (1080 x 1350), optimized for Instagram feed. - Compose each image so it also crops cleanly into a square for Instagram grid use. - Keep Don centered enough for square-crop safety. - Use the exact same fictionalized Don Scott character in every image: same face, hair, age, build, skin tone, and overall appearance. - Style should feel like realistic civic campaign photography in Bedford, New York: warm, local, natural light, lightly staged but believable. - Humor should come from the concept, not from making the image visually absurd. DO NOT INCLUDE - Any text in the image - Logos or campaign branding - Watermarks - Readable political signs or readable signage - Partisan symbols - Giant flags - Distorted hands - Weird AI artifacts - An exaggerated politician grin IMAGE REQUEST Generate a realistic civic campaign-style photo of the same fictionalized Don Scott character for the Bedford Town Supervisor UNCAMPAIGN. SCENE Don is standing outside a small local community building in Bedford, New York, at a modest ribbon-cutting ceremony. He is holding oversized ceremonial scissors near a ribbon, with a few local residents nearby. The mood is warm, neighborly, and slightly staged in the familiar political photo-op way, but restrained and believable. Natural daylight. Small-town suburban Westchester feel. VISUAL STYLE - Realistic campaign photography - Warm, clean natural light - Lightly polished editorial look - Calm, modest smile - Practical local-candidate clothing