Braven’s Scale of AI use

Four levels that say how much of a project was made with AI, from a little help to work run entirely by agents.

Why I use AI, and why I say so

In my opinion, AI is a revolutionary technology that allows unprecedented, rapid development in many fields. Instead of gating and limiting its use as most institutions do, I let myself use AI tools to their full potential and push the limits of what they can do. That way I keep learning about this ever-changing, rapidly improving technology, and it happens to speed up my work and raise its quality.

These tools add a layer of abstraction between knowledge and understanding. I believe that layer is what lets us focus on the more important goal of a project.

However, I am very much against using AI tools to make low-quality, low-effort work, especially art. I use AI tools in my design work, but I heavily tweak the design styles myself until they reach a level I would be proud to display.

That is why I think it is important to be open about how I use AI in all of my projects and work, and why I created this scale.

The four levels

Each level includes everything in the levels before it.

  1. L1 mark: little use of AI based tools

    Little to insignificant use of AI

    I used AI tools in the project, but they had little to no impact on the outcome.

    • Paraphrasing
    • Ideas
    • Feasibility checks
    • Small, minor pieces of work
    • Generating charts
    • Checking my work
  2. L2 mark: average use of AI based tools

    Moderate use of AI

    Everything in Level 1, plus AI tools were used to create or generate core parts of the project.

    • Writing code
    • Assets the project depends on, like textures for a 3D object
    • Configuring devices
  3. L3 mark: heavy use of AI based tools

    Heavy use of AI

    Level 3 takes this one step further. AI tools ran in automation loops and made crucial, impactful decisions for the project on their own, without asking a human first. The code is almost entirely written by AI, with little to no human involvement.

    Humans take a higher-level role: they orchestrate and manage the work the AI agents produce. Humans are still in the loop and make the key decisions.

    • Agents in automation loops
    • Almost all code written by AI
    • Humans orchestrate and review
  4. L4 mark: entirely created by AI tools

    Entirely AI generated

    At Level 4 the orchestration role passes to an agent. Humans are no longer in the loop, and agents make every decision on their own.

    A human sets the goals and leaves the agents running in a loop. Human intervention is rare and happens only when needed. The results reach the human to use or study only after the work is complete.

    • A human sets the goal
    • Agents orchestrate other agents
    • Results reviewed after completion

The levels side by side

The same questions, answered for each level.

Comparison of the four levels of AI use
L1Little to insignificant use of AILittle useL2Moderate use of AIAverage useL3Heavy use of AIHeavy useL4Entirely AI generatedEntirely AI
How much AILittle to insignificantModerateHeavyEntirely AI generated
What AI works onSmall supporting tasksCore parts of the projectAlmost all of the work, in automation loopsAll of the work
Who does the core workMeMe, with AI generating core partsAI agentsAI agents
Who makes the key decisionsMeMeMe, while AI makes some crucial calls on its ownAgents, working toward goals I set
AI acts without askingNoNoYesYes
Human in the loopYes, throughoutYes, throughoutYes, for the key decisionsNo, only at the start and the end
My roleMakerMaker, using AI as a toolOrches­trator and managerSets the goal, steps in rarely, studies the results
Example usesPara­phrasing, feasibility checks, charts, checking my workWriting code, 3D textures, configuring devicesAgent loops that write almost all of the codeAgents running on their own toward a set goal

The QR code on each printed mark links back to this page.

© Chiambucket 2025
Credits to content