Question 1

Does the AI Literacy Evidence Kit certify that we comply with the EU AI Act?

Accepted Answer

No. It builds the audit-ready evidence record — role by role — and tells you whether that record is defensible, but it does not certify compliance and it is not legal advice. The EU AI Act Article 4 AI-literacy obligation has applied since Feb 2025, with penalty and governance provisions phasing in from Aug 2025, and the dates move as the Commission's Omnibus simplification package settles. Confirm your obligations with qualified counsel.

Question 2

How does the Kit decide our organization is NOT DEFENSIBLE?

Accepted Answer

It takes your weakest in-scope role as the headline, not the average. Each role scores 0-100 across four evidence levers (inventory, role mapping, training, proportionality, weights summing to 100) and lands EVIDENCED, PARTIAL, or UNEVIDENCED. The org reads DEFENSIBLE only if every in-scope role is EVIDENCED; SHORTFALL if the weakest is PARTIAL; NOT DEFENSIBLE if any in-scope role is UNEVIDENCED. One unevidenced operating role is a defensibility hole, so a 90%-covered org with one such role still reads NOT DEFENSIBLE.

Question 3

What is the gate in the AI Literacy Evidence Kit?

Accepted Answer

Any in-scope role whose staff operate an AI system with zero training record is UNEVIDENCED on its own and forces the whole organization to NOT DEFENSIBLE — regardless of every other score. That untrained operator is the exact exposure Article 4's proportionality duty is about. The gate clears the moment a training record exists for that role, so the fastest move out of NOT DEFENSIBLE is usually to train and record the operating roles first.

Question 4

How is this different from the AI Literacy & Workforce Training Kit?

Accepted Answer

This is the evidence gate that runs before and around the training. The AI Literacy & Workforce Training Kit delivers the program — the deck, workbook, knowledge check, and completion log. This Kit reads your tool inventory, role-to-system mapping, and training records and scores whether your program is proportionate and defensible. Prove it, then train: run this first to find the holes, run the Training Kit to fill them, then re-run this to confirm the record holds.

Question 5

Does the Kit need to connect to our systems or upload our data?

Accepted Answer

No. It is a runnable Python engine plus a workbook that reproduces it exactly — you enter your own roles and ratings, and it computes locally. Nothing is uploaded and nothing connects to a live system. The numbers are yours; the engine never invents a rating or a training record, and it is deliberately date-agnostic so it never treats a shifting deadline as the verdict.

Question 6

Who is the AI Literacy Evidence Kit for?

Accepted Answer

It is for the person who owns AI governance or L&D in an organization that uses AI and may fall under the EU AI Act — a founder, an operations or compliance lead, or an L&D owner. If an auditor, insurer, or enterprise buyer asked you today to show your AI-literacy program is proportionate per role, this is the artifact that answers them. It pairs with the NIST AI RMF Readiness Kit and the EU AI Act Readiness Kit for the wider governance picture.

Prove your AI-literacy programis defensible — before you train.

“We trained the team” is an assertion. An auditor wants the record.

Edit the ratings — the org verdict tracks your weakest in-scope role, not the average.

The same scoring runs as a Python engine and a workbook that reproduces it exactly.

Three operating principles keep the evidence read honest.

An evidence record you can show. Not a compliance certificate.

The person who'd have to answer if an auditor asked to see the evidence.

Run this first; it routes you to the program and the wider framework.

Straight answers on scope, the gate, and what this does and doesn't certify.

Prove it, then train.
Find the hole before the auditor does.