Skip to Content

No AI Slop Manifesto: five values for working with AI output

What the manifesto says about AI slop, and which team agreement to take from it today.
November 17, 2026

Speakers and attendees of the AIE Code Summit 2026 wrote the No AI Slop Manifesto: five values and a handful of resolutions for working with AI output. It is not a law or a standard, but a conversation piece for the workplace. That is exactly why it is useful: with your team, you can take one working agreement from it on who reviews AI output and when you say an agent wrote it.

What slop is, according to the manifesto

The definition is sharp: slop is when we put less effort into producing something than we ask the reader to spend understanding it. So using AI does not automatically make something slop. It becomes slop when you pass AI output along without taking responsibility for the quality and for the reader.

Two points from the manifesto affect every SME team:

  • Slop is asymmetric. A ten-page document costs almost nothing to produce, but reviewing it still takes just as long.
  • Slop compounds. The more junk sits in your knowledge base, chat and code, the more of it the next agent reads and copies.

Our view: that second point is underestimated. Anyone working with a second brain or agents notices that a sloppy note today pollutes tomorrow's answer.

Five values, modelled on the Agile Manifesto

The authors refer to the Agile Manifesto of 2001, with which seventeen software engineers and leaders changed how the industry builds software. This manifesto chooses five values:

  • Outcomes over output
  • Understanding over speed
  • Our own thinking over a model's first draft
  • Better work over easier work
  • Improving the system over patching the output

Agreements that apply directly to client work

A few of the resolutions stand out for anyone who sends AI text to the outside world:

  • Say so. If an agent wrote something you are sending, make that clear and check the parts that matter before it goes out.
  • Speak up, kindly. If you receive slop, say so, and ask the sender to review it themselves first.
  • Fix it in the system. If you see the same mistake twice, improve the process or the check, not just the text.

How much you review depends, according to the manifesto, on what is at stake. Not everything needs a human.

Our view: the line is the client. Whatever goes out is read by a person who can judge its content.

What to do now

  • Read the manifesto with your team, in a half-hour meeting.
  • Make one working agreement for AI text that goes to clients: who reviews it, and when do you mention that an agent wrote it. Start with quotes, contracts and client emails.
  • Keep track of mistakes that come back twice and adjust your prompt, template or check accordingly.

How it works under the hood

The manifesto is not a technical specification. It does say: whatever a linter, type checker or other automated check can catch, let it catch. Agent reviews and automated checks are useful, but according to the authors not sufficient. For now, a human still reviews at least part of the output, depending on the risk. The manifesto is open for signatures and comments; on 17 November 2026 it had 714 signatures.

This news touches chapters 20 and 17 of the course AI(l) or nothing by Newway. If you follow the course, this subject returns there at a calmer pace.

What you can do with this

Want to know what this means for your business? Plan an introduction: half an hour, no sales pitch, but an honest answer to the question where to begin. Rather read on first? The course AI(l) or nothing is free and starts with the basics.

Written in response to publications by Latent Space and No AI Slop: A Manifesto.