8000 uail/SPEC.md at v1.0 · simon-inkie/uail · GitHub
[go: up one dir, main page]

Skip to content

Latest commit

 

History

History
77 lines (48 loc) · 8.34 KB

SPEC.md

File metadata and controls

77 lines (48 loc) · 8.34 KB

UAIL: the Universal AI Level

Version: v1.0 Published: August 2026 Official source: Inkie Live page: inkie.ink/universal-ai-level

This file is the canonical text of the spec. Level names, descriptions, "In plain English" statements and "How do you know?" tests are normative: don't reword them, quote or extend instead.

Definition

The Universal AI Level, or UAIL, is a practical way to describe someone's current ability to use artificial intelligence.

Your UAIL is the highest level at which you can work independently and repeatedly.

It is based on what you can reliably do, rather than which tools you have opened, which terminology you recognise or how confident you feel when talking about AI.

Using a feature once does not necessarily place you at that level. You should be able to repeat the task, understand the result, recognise common problems and make reasonable progress when something goes wrong.

AI Literacy Levels: 0-10

Level Name In plain English What you can do How do you know?
0 Uninitiated "I haven't used AI." You have never intentionally used an AI chatbot, image generator or similar AI tool. You may use products containing AI without realising it, but you have not directly instructed an AI system yourself. You would need help choosing a tool, opening it and beginning your first interaction.
1 Conversational "I talk to AI." You can hold a useful conversation with an AI assistant. You can ask questions, provide context, request writing, give feedback and refine an answer. When the first response is unsuitable, you can explain what needs changing and continue the conversation without starting again.
2 Creator "I use AI to make things." You can use AI to create different types of output, including images, documents, presentations, spreadsheets, audio or video. You understand the basic process of generating, reviewing, editing and using an output. You can begin with an idea and produce something usable, rather than copying the first answer without reviewing it.
3 Operator "I use AI confidently in my work." You can use AI for meaningful work involving research, analysis, planning, source documents and data. You can choose suitable tools, provide useful context, improve instructions and check the quality of the result. When an answer, formula or conclusion looks wrong, you investigate it. You do not assume that a confident AI response must be correct.
4 Automator "I create repeatable AI workflows." You can design custom assistants, agents, templates and automations that complete repeatable tasks. You can structure inputs and outputs, pass information between stages and improve a workflow when it produces inconsistent results. You can design a useful process from scratch rather than relying entirely on a ready-made template. When it fails, you can usually identify which stage caused the problem.
5 No-Code Builder "I build connected AI systems without coding." You can build complete no-code or low-code applications where AI tools and other services work together. You are confident using connectors, MCP, APIs through visual tools, databases, authentication and agent systems without writing the underlying code yourself. You can build and maintain a working system involving multiple tools, data sources or agents. When a connection breaks, you know where to look and can make meaningful progress towards fixing it.
6 AI-Assisted Coder "AI helps me code." You can use AI to write and edit code, create scripts, work with repositories, use a terminal, connect APIs and deploy a simple application. You may still rely heavily on AI to explain the code and solve technical problems. You can take a simple software idea from instruction to working deployment, although unfamiliar errors may still leave you dependent on AI or another developer.
7 AI Developer "I build and maintain AI applications." You can independently inspect, change, test and debug the code behind an AI application. You understand databases, authentication, APIs, application architecture, security and deployment well enough to maintain what you build. When AI-generated code fails, you can read it, identify the likely cause and make informed changes rather than repeatedly asking the AI to try again.
8 AI Systems Engineer "I architect production AI systems." You can design reliable systems involving models, custom agent harnesses and loops, retrieval, orchestration, evaluation, monitoring and model routing. You can balance accuracy, speed, security, reliability and cost. You design for failures, misuse, changing models and unexpected inputs before customers discover them for you.
9 AI Architect or Researcher "I build or adapt the underlying AI." You can train, fine-tune or substantially adapt models. You understand training data, evaluation, optimisation, infrastructure, safety and machine learning at a deep technical level. You can develop or materially improve the models, evaluation methods or infrastructure that other AI developers build upon.
10 Creating the Matrix "I created the AI." You develop foundational models, architectures, protocols or genuinely new approaches to artificial intelligence. Your work changes what other researchers, engineers and organisations are able to build. Other AI experts use your research or technology to create their own systems. You may describe the work as "fairly straightforward" immediately before drawing on six whiteboards.

The code line

UAIL 0-5 can be reached without coding. These levels cover conversation, content creation, professional AI use, workflows, automation, connected systems and no-code application building.

UAIL 6-10 involve code. These levels cover AI-assisted software development, application engineering, production architecture, model development and foundational research.

This distinction matters because an AI course can be described as suitable for beginners while quietly expecting learners to understand Python, APIs, repositories and the command line. UAIL makes that expectation visible.

Broad capability groups (the four colour bands)

  • UAIL 0-2: AI User, you are beginning to use AI through conversation and creative tools.
  • UAIL 3-5: AI Practitioner and No-Code Builder, you use AI independently in real work and can create increasingly sophisticated workflows, automations and connected systems.
  • UAIL 6-9: Developer, Engineer or Architect, you use code to create, maintain and improve AI-powered applications, infrastructure and models.
  • UAIL 10: Foundational Creator, you are developing the technologies and approaches that shape what artificial intelligence can become.

How course creators can use UAIL

Courses should state both the expected starting level and the intended outcome.

Example: "This course is designed for learners at UAIL 2-3. No coding is required. By the end of the course, learners should be able to work independently at UAIL 4."

This is more useful than describing a course as beginner, intermediate or advanced because it explains what learners should already be able to do and what practical capability they should gain.

A course should not promise that someone will reach a particular UAIL unless its teaching and assessment cover the full capability described at that level.

How employers can use UAIL

Employers can use UAIL to describe the practical AI capability required for a role.

Example: "This role requires approximately UAIL 4. Candidates should be able to design and maintain repeatable AI workflows without coding."

UAIL should support clear role descriptions rather than become an arbitrary barrier. Someone does not need to know every available platform to demonstrate that they can work at a particular level.

How individuals can use UAIL

Your UAIL is a description of your current capability, not a score for your intelligence, value or future potential.

You may also have different levels in different areas. Someone might be highly capable at AI-assisted research but have little experience with image generation or application development.

Use your overall UAIL to describe the highest level at which you can work reliably, then add specific strengths where useful.

Example: "UAIL 4, with particular experience in marketing automation, research and custom assistants."

0