AI Influence

How much of each post here an AI wrote, and how the posts get made.

Posts on this blog carry an AI Influence Level badge next to their tags, and a note at the end that says the same thing in words. The scale is Daniel Miessler's AI Influence Level (AIL), used here at version 2.0, exactly as he defines it.

To pick a level, read down the list and stop at the first line that's true of the post. His levels and lines:

  1. AIL 0: Human Created, No AI Involved. Nothing in the text came from an AI.
  2. AIL 1: Human Created, Minor AI Assistance. An AI only tidied the text: proofreading, a grammar fix, a transcript cleaned up.
  3. AIL 2: Human Created, Major AI Augmentation. An AI added to text a human wrote, rather than just tidying it.
  4. AIL 3: AI Created, Human Full Structure. An AI wrote it inside a structure a human fully laid out.
  5. AIL 4: AI Created, Human Basic Idea. An AI wrote it from a basic idea a human brought.
  6. AIL 5: AI Created, Human Request. A human asked for it, and the AI came up with the idea itself.
  7. AIL 6: AI Initiated, No Human Involvement. Nobody asked. The AI decided the piece should exist and did everything after that.

Levels and lines from The AI Influence Level (AIL) by Daniel Miessler, version 2.0 (September 2026).

AIL rates the text. Art is rated on its own, so a post's note gives its featured image a separate level.

I run an AI agent, the same kind of agent I help clients put into production safely. It does most of the writing on this blog, and the badge is how I say so.

  • Essays. Most start as a topic the agent proposes from the security news it reads every day. I pick which ones get written and sometimes steer the angle. The agent drafts the post in our CMS, and I decide when it goes live. That's AIL 5. When the idea is mine, it's AIL 4.
  • The Tuesday Briefing. An AI pipeline writes each edition from the week's news, on a schedule I set up, and it publishes without my review. That's AIL 5.

A level tells you how a post was made, and nothing else. I'm accountable for every post either way. If one gets something wrong, tell me.