For Politicians and Elected Officials

How Politicians Can Use AI Without Losing Authentic Voice and Voter Trust

AI can write a grammatically perfect speech that your constituents will never believe came from you. Tools like Claude and ChatGPT can generate policy positions so coherent that you forget they were not tested against the real concerns you hear in town halls. The risk is not that these tools are bad at their job. The risk is that they are so good at sounding like authority that they can replace the actual work of forming judgement through argument, community knowledge, and deliberation that voters expect from their representatives.

These are suggestions. Your situation will differ. Use what is useful.

Download printable PDF

Use AI to Draft, Not to Replace Your Thinking on Policy

When you run a policy proposal through Perplexity or Claude to find supporting evidence, you are outsourcing your research but not your reasoning. The danger comes when you adopt the AI's framing of the issue as your own position without testing it against what you know from your local context, your constituents' actual lived experience, and your own values. Policy positions that emerge from AI analysis alone will lack the conviction that comes from honest disagreement and argument. Use the AI output as a starting point for the deliberative work you should be doing anyway: asking whether this position fits your community's priorities, whether it survives scrutiny from people who disagree with you, and whether you could defend it in a room full of sceptical constituents.

›Run AI-generated policy summaries past three people in your constituency who disagree with you before you adopt the position publicly
›Ask yourself whether you could explain the core logic of the position without reference to the AI output
›When you use Quorum AI for legislative research, treat its analysis as evidence to interrogate, not as the conclusion

Write Your Own Speech Openings and Closings

AI can assemble a competent speech body from your talking points and policy briefs. What it cannot do is capture the moment that made you run for office, the specific family story that drove your commitment to a particular cause, or the authentic connection you feel to the place you represent. Constituents hear the difference between a speech that begins with a real memory and one that opens with a plausible-sounding anecdote crafted by ChatGPT. The sections of your speech that matter most for trust and authenticity are exactly the sections you need to write yourself. Use AI to help you structure your argument or fill in background detail. Write the parts that only you can write.

›Draft your opening and closing in your own voice before you give anything to Claude or ChatGPT
›If you use AI to flesh out the middle sections, read the full speech aloud to yourself and rewrite any paragraph that does not sound like you speaking to a room
›For high-stakes speeches, read your own draft to a colleague and ask whether they can hear your voice

Personalised Constituent Communication Must Reflect Real Relationships

Scaling personalised emails to constituents using AI tools creates a genuine problem for electoral trust. When voters discover that their heartfelt message received a response generated by an algorithm, it reveals that the relationship they thought existed with their representative was a computational fiction. This is not just a reputational risk. It is a damage to the social foundation that representative democracy rests on. You can use AI to help you process the volume of constituent mail you receive. You cannot use it to replace your own attention to the concerns that matter most to people in your constituency.

›Reply personally to messages on issues that directly affect your policy decisions or involve constituents you know
›If you use Claude or ChatGPT to draft responses to routine enquiries, include a line that makes clear which communications receive your personal attention
›When constituent feedback shows a strong local pattern on a single issue, that is a signal to read several messages yourself and form your own view before responding

Media Strategy Should Amplify Your Judgement, Not Replace It

AI can analyse media coverage and suggest talking points with impressive speed. Microsoft Copilot can help you spot patterns in how your positions are being reported. The risk is that you begin to let the AI's assessment of what is newsworthy or what matters override your own political judgement about which issues deserve your attention. You know your constituency better than any algorithm. You understand which fights matter to the people who voted for you and which ones are distractions. Use AI to help you see what the media ecosystem is doing. Use your own judgement to decide what you should do in response.

›When Perplexity or Claude identifies a media narrative, ask yourself whether you agree that it reflects what actually matters to your constituents
›Do not let AI-generated talking points become your primary message. Use them as one input alongside direct constituent feedback
›Before you respond to a media story, ask whether responding serves your constituents or whether it mainly serves the story

Build Transparency Into Your AI Use Before Someone Else Exposes It

Political figures who use AI without acknowledging it face credibility collapse when the use becomes public. Voters are increasingly sceptical of communications that feel constructed rather than authentic. Transparency about where you use AI and where you do not is a form of honesty that builds trust rather than eroding it. This does not mean announcing every use of Copilot or Claude. It means being clear with your constituents about the role of technology in your communication and ensuring that the work that requires your personal judgement visibly comes from you.

›If you publish policy research generated using AI, note the AI tools you used and what additional human analysis you applied
›Be prepared to explain to journalists why you used AI for particular tasks and why you did not use it for others
›When a constituent asks whether something was written by you or by AI, answer directly

Key principles

1.AI should compress research and drafting work, never replace the deliberation that creates genuine conviction in your policy positions.
2.The parts of your communication that build voter trust are the parts that only you can produce: your own voice, your real stories, your actual judgment.
3.When constituent communication at scale is indistinguishable from algorithmic output, you have replaced the relationship that electoral trust depends on.
4.Political authenticity is not incidental to your effectiveness. It is the core asset that allows you to lead rather than simply manage.
5.Transparency about where you use AI is more credible than hiding it, and it gives you control of the narrative before someone else defines it.

Key reminders

When you use ChatGPT or Claude for speechwriting, write the first draft yourself so the AI output builds on your voice rather than replacing it
Test policy positions generated by AI against the three most common objections you hear from constituents who disagree with you
For constituent mail, create two categories: messages requiring your personal response and messages where AI-drafted responses are appropriate, then be honest about the distinction
Ask your communications team to flag any statement or position that came entirely from AI analysis before it goes public
Every quarter, review which parts of your work now come from AI and which still require your personal judgment, then consciously protect the latter