Artificial Intelligence Quotient

How well do you actually use AI?

AI systems excel at some surprisingly complex tasks while failing at ones that seem simple. That irregular, ever-shifting boundary is the jagged edge. AIQ measures how well you read it — and how well you put it to work.

Take the AIQ test Learn more

The concept

AIQ is a human's ability to understand and productively use the jagged edge of AI capability.

High-AIQ people know where the AI boundary is today, choose tasks and prompts that align with current strengths, and apply the right human checks in areas of weakness.

Low-AIQ people tend toward one of two failure modes: they either dismiss AI altogether, or they over-trust a hallucinating, sycophantic system with decisions it shouldn't make alone.

Unlike IQ, a relatively stable trait, AIQ is a perishable skill. Model updates arrive weekly. Yesterday's reliable shortcut may be today's failure point. High AIQ requires active maintenance.

Read the original essay: AIQ: Rethinking Human–AI Capability.

Five dimensions, scored 0–100

Updated with the frontier
01 · Prompting
Crafting clear instructions
Writing adaptive prompts that align with what the model is actually good at right now — not what you assume it can do.
0 – 20 pts
02 · Context Awareness
Knowing what the AI has and needs
Understanding what information is in the model's context window, what it lacks, and what that means for the reliability of any output.
0 – 20 pts
03 · Anticipating Failures
Recognising hallucination and sycophancy
Spotting when a model is confidently wrong, agreeing to please rather than inform, or operating outside the zone where it's trustworthy.
0 – 20 pts
04 · Verification & Ethics
Applying human judgment and guardrails
Knowing when to verify, how to verify, and where human oversight is not optional — especially in health, finance, education, and safety.
0 – 20 pts
05 · Workflow Integration
Embedding AI in processes that hold
Building reliable human–AI workflows that convert capability into consistent, auditable output — not one-off wins that can't be reproduced.
0 – 20 pts
0–100 Your AIQ score

AIQ assessments must be re-baselined regularly as the frontier shifts. Items retire, new ones are added. An AIQ credential that never expires is one that cannot be trusted. The test you pass today is not the test you'll need to pass in six months — and that's precisely the point.

Research

Ten independent teams. One shared intuition. No one was talking to each other.

In a survey of ten independently developed AIQ frameworks across cognitive psychology, computer science, education, enterprise consulting, and marketing, a clear pattern emerges: most measure knowledge or performance while missing the central mechanism — calibration.

The paper maps the landscape, introduces a two-dimensional analytical model, and argues that AIQ must be treated as a perishable skill assessed against a moving frontier — not a stable trait.

Read the paper →

Find out how well you really use AI.

A 25-question self-assessment across five dimensions. Takes about 8 minutes. You get a scored breakdown and a radar chart — no signup required. This is a preliminary orientation test; a fuller open-ended assessment is on its way.

Take the test — it's free

Developed by Wim Sweldens · Read the research on Substack