Substack's new AI-detection tool is a quiet admission that the platform is no longer willing to let readers guess. By giving audiences a way to estimate how much of a newsletter was written by artificial intelligence, Substack is trading the comfortable ambiguity of "assisted" for something closer to a ledger. That is a meaningful step, and it deserves more than a shrug. We have spent the last year watching AI tools blur the line between drafting and writing, and the result is that trust has become the scarcest resource in publishing. This move is not about policing creativity; it is about giving readers a baseline so they can decide what they actually want from the writers they follow.
For you, the reader, this changes the relationship you have with the newsletters you open every morning. The tool does not tell you whether the AI output was good, or whether the author edited it heavily, or whether the idea was original. It gives you a percentage, and percentages are blunt instruments. But blunt can still be useful. If you are someone who subscribes to a newsletter because you value a human perspective, a high AI estimate is a prompt to ask a better question: not "did they use AI?" but "what did they use it for?" That is the practical shift here. It moves the conversation from fear to discernment, from vague anxiety to a concrete evaluation of process. We would tell you to treat this feature as a starting point, not a verdict. A writer who uses AI to brainstorm an outline and then rewrites everything by hand is not the same as one who pastes a prompt and publishes the output. The tool cannot see that difference, but now you can look at the number and then read the work with more context.
What we find genuinely interesting is what this signals for the broader ecosystem. If Substack is willing to surface this metric, expect other platforms to follow. Not because they are all committed to transparency, but because readers are starting to demand it. The genie is out of the bottle, and pretending otherwise is no longer a strategy. This is where we land: the tool is not a moral high ground, it is a mirror. It reflects the messy reality that most writing now exists on a spectrum. Our take is simple. If you are a writer, do not hide from the metric. Use it to have an honest conversation with your audience about your workflow. If you are a reader, do not use it to shame or dismiss. Use it to make an informed choice about where you spend your attention. The real question Substack has just put on the table is not whether AI wrote the words. It is whether you trust the person who put their name on them. That is a question no percentage can answer, but at least now you can ask it with your eyes open.