Getting Started with AI: FAQs
Plain-language answers for campaign staff, organizers, and progressive practitioners: where to start with AI, how to stay safe, what it costs, and how to connect with others.
Getting started
How do I get started in an intentional way?
Start small: choose one predictable, text-rich workflow with clear inputs and outputs, such as event follow-up emails, meeting-note tasks, a translation, or a long-document summary. Try it for two weeks, improve it, then decide whether to expand.
What kinds of AI show up in practice?
Most campaign and organizing work falls into four categories: large language models for drafting and analysis, agentic systems that take actions, AI built into existing software, and purpose-built campaign tools.
Does it matter which model I use?
Yes, but there is no permanent winner. Larger models tend to help with long documents, nuanced writing, and multi-step reasoning; smaller models are often enough for fast drafts and summaries. Practice and careful review matter more than allegiance to one provider.
Using AI well
What are foundational tools to start with?
Audit AI features in your current tools, then try a conversational tool such as Claude, ChatGPT, or Gemini, a meeting-notes workflow, a scheduling tool, and a simple document-summarization prompt before buying new software.
What does this cost?
Free tiers are useful for learning and light work. Paid plans commonly cost about $20 per user per month and help with longer documents and stronger models. If staff will handle voter data, donor information, or internal strategy, use an appropriate business, enterprise, or API tier first.
How do I write a good prompt?
Give clear instructions, relevant context, a desired format, and an example when possible. Ask for a specific output—such as a 150-word fundraising email for a named audience—rather than a vague request.
Where can I start integrating AI in my work?
Look for tasks that are repetitive, text-rich, and reviewed by a person before the result matters. Use caution where being wrong has serious consequences or where human trust, political judgment, and lived experience are central.
Safety, oversight, and policy
How can I engage more securely with AI tools?
Do not paste voter PII, donor financial data, or sensitive strategy into consumer tools without reviewing data practices. Give connected tools the least access they need, read the output before it goes public, and build an AI use policy covering approved tools, data rules, review, and reporting.
Where does human oversight fit?
Human judgment belongs at every consequential step: people prompt, review, edit, approve, and remain responsible for public-facing, legally significant, politically sensitive, and relational work.
How do I create an AI use policy?
A useful policy names approved tools, data that may and may not be submitted, required review, transparency norms, and how staff raise concerns. Keep it short enough to use, and revisit it every six months.
How can I reduce environmental impact?
Prefer text when it will do the job, choose appropriately sized models, batch requests, and consider the energy and governance effects of the larger data-center buildout—not only the footprint of a single prompt.
Learn and connect
Where can I find more tools and training?
Browse the tools directory, Progressive AI Skills, formal training resources, and the workflow library. These resources are suggestions rather than endorsements; choose tools that fit your workflow, budget, and data practices.
How can I connect with other progressives thinking carefully about AI?
Join an AI Open Mic, explore ecosystem writing and training in the resource library, and share questions or working examples with the community.