1,017 AI skill assessments in. the data is humbling.

crossed 1,000 assessments on aisa.to last month. here's the thing that keeps bugging me:

average score is 52 out of 100. but the distribution isn't what you'd expect — it's not a normal curve. there's a huge cluster in the 40-55 range (we call them Enthusiasts) and then almost nobody above 80.

out of 1,017 people, zero scored at Oracle level. not one.

the gap isn't knowledge. most people can name 5+ AI tools. the gap is verification — do you actually check whether AI output is correct before using it? that single behavior separates the top 15% from everyone else.

we published the full breakdown here: aisa.to/state-of-ai-fluency

building an AI assessment tool in a market where everyone thinks they're already good at AI is... an interesting positioning challenge. but the data keeps being consistent — and that's what keeps me going.