A reader told me my AI-assisted essay didn't sound like me anymore. That critique kicked off a week where reputation and AI collided in public three times: Substack's new Pangram "human-detection" scan, LinkedIn's AI-slop report button, and the backlash over an influencer's OpenAI brand trip. Each one exposes the terms platforms are quietly setting for anyone with a public body of work. This episode is me working out how to stay visible without giving up my own ground.
Pillars: Trust · Visibility · Identity · Communication
What's covered
A critique that stuck (00:00–03:00) A reader told me my AI-assisted essay had lost my voice — that it sounded more like Claude than like me. I disagreed. But the position got harder to defend once I separated two questions: can I write, and why am I experimenting at all.
Where AI actually fits my work (03:00–04:00) Useful for SEO, AEO, and discoverability. The trainer material — takeaways, reflection questions, anything designed to teach rather than model — is mine to write. That part was never up for experiment.
Substack + Pangram (03:52–05:41) Substack now scans posts and rates them human, AI-assisted, or AI-written. My read: unreliable like every detector, and easy to game, since anything under 100 words, audio, video, or paywalled content can't be scanned at all. So who benefits? Substack gets a PR win. Pangram gets a funding-round talking point. Writers get scanned. My response: I'm moving my experimentation behind the paywall.
LinkedIn's AI-slop button (05:57–07:12) No test, no proof, just crowd-sourced suspicion anyone can act on. Any anonymous reporting system carries the same abuse potential, and we already know where that lands: ageism, misogyny, racism.
Thought leader vs. influencer, and the OpenAI brand trip (07:12–10:54) A creator I follow — she wouldn't call herself an influencer, but she functions as one — took an unpaid OpenAI brand trip and posted about it while running a bath. The comments tore her apart over data-center water use. Her defense was "I didn't know this would be controversial." That's a research failure, and for someone whose credibility rests on thinking things through, it's a costly one.
Trust, reputation, and PR strategy (10:54–12:27) Getting torn apart in the comments isn't the answer either. I think we're capable of civil disagreement on both sides. But if you're a public persona doing brand deals, you need a real strategy: know your stuff, own it, explain it, and make sure the deal's values match yours.
Platform incentives (12:34–15:00) LinkedIn's new multi-collaborator post feature runs on the same engagement logic under a friendlier name. More talking helps the platform at least as much as it helps you. And this goes past creators — if you're an employee, your public presence is also your resume.
Reflection questions
- Is it actually worthwhile to take the risk of experimenting publicly with AI?
- What does a platform gain by having users rate each other as "less than human"? Who does that serve?
- Am I experimenting with this tool because it's genuinely useful, or just because it's there?
- Whose credibility is most at risk when anonymous, crowd-sourced flagging gets abused?
- Where's the line between thought leadership and influence, and does that line change how I'm judged?
- What is it about a voice I don't fully agree with that still works? What can I learn from that discomfort?
- Before taking a partnership or brand deal, have I actually researched who I'd be standing next to?
- Do public missteps deserve public dismantling, or is there a more civil way to hold the same concerns?
- How do I present myself in a way that builds trust and protects my reputation without getting defensive?
- Would I want my name permanently tied to this company or platform, in this form, five years from now?
- Am I transparent about how I'm using AI? Does that use actually serve me, my work, and my relationships?
- What future is this choice building toward, not just what does it do for me right now?
- How do I stay visible in a way that still feels ethical to me?
A note on process: these show notes are pulled from my own transcript, not generated from a prompt. AI did the structuring and compression; the words are mine. Same distinction I draw in the episode: extraction and discoverability work, not voice work.
It's not safe to be experimental for free anymore. I'll still be on LinkedIn, and the longer essays will still land on Substack — but the experiments are moving behind the paywall. If you want to watch the process, that's where it lives now.