← All lessons
Level 1 · Aware6 min read

What AI compliance is (and who the rules target)

AI compliance is the routine work of making sure the AI your organisation builds or uses stays inside the law, matches what you tell people it does, and can be explained when someone asks. It is not a single certificate. It is a set of habits: knowing what you use, judging how risky each use is, keeping the right records, and giving people the notices and choices the law requires.

Two roles the law cares about

Almost every AI rule sorts you into one of two roles, and many organisations are both at once.

  • Builder (the EU AI Act calls this the "provider"; Colorado and Texas say "developer"): you train, fine-tune or substantially modify a model or system, or you put one on the market under your own name. Builders carry the heaviest duties: technical documentation, testing, disclosures to customers, and in some cases registration.
  • Deployer (also "user" or "employer" in some state laws): you use an AI system in your own operations, for example a hiring screen, a support chatbot or a credit model. Deployers must use the system as instructed, tell affected people, keep logs, and keep a human in the loop where the law says so.

If you take a vendor's model and wrap it in your own product, retrain it on your data, or change its intended purpose, several laws treat you as a builder of the new system. The EU AI Act says this explicitly for anyone who makes a "substantial modification" or rebrands a high-risk system (Article 25).

What compliance actually covers

Most frameworks, whether the EU AI Act, the NIST AI Risk Management Framework or ISO/IEC 42001, touch the same ten areas. They are the ten domains of the assessment in this Academy:

  • Inventory: a list of every AI tool and model in use.
  • Risk classification: which uses are prohibited, high-risk, or low-risk.
  • Data and privacy: what personal data goes in, and on what legal basis.
  • Transparency: telling people when AI is involved or content is synthetic.
  • Human oversight: a person who can check, override or stop the system.
  • Fairness: testing for discriminatory outcomes.
  • Security and robustness: resisting attacks and behaving predictably.
  • Vendors and contracts: knowing what your suppliers promise and prove.
  • Governance: a named owner, a policy, and a way to handle incidents.
  • Training: staff who understand the tools they use.

Why the order matters

You cannot classify risk without an inventory, and you cannot prove oversight without records. That is why the assessment caps your maturity level at "Emerging" until inventory and risk classification are in reasonable shape. Start there; the rest builds on it.

A note on advice

This Academy explains rules in plain words and points you to official sources. It is general information, not legal advice. Laws change, they differ by place and by sector, and how they apply depends on facts only you and your counsel know. Use the lessons to ask better questions, then confirm the answers that matter with a qualified professional.

What this means for you

Decide today whether you are a builder, a deployer, or both, for each AI use you know about. That single decision determines most of the duties in every later lesson. If you are unsure, assume the stricter role until you have checked.

Real case

Lawsuit2025 · US-federal

Mobley v. Workday — collective action over AI hiring screens (ongoing)

Ongoing as of September 2026. In July 2024 the court refused to dismiss the case, accepting the agent theory. On 16 May 2025 it conditionally certified a nationwide ADEA collective of applicants aged 40+ rejected via Workday's platform since September 2020; the opt-in notice period closed on 7 March 2026. In June 2026 the court let California FEHA claims and a proxy-discrimination disability claim proceed. No trial date, judgment or settlement has been reported.

Read the case →

Related laws

Quick check · 3 questions

  1. 1.Under the EU AI Act, what does the law call the organisation that develops an AI system or puts it on the market under its own name?

  2. 2.You license a vendor's hiring model, retrain it on your own applicant data and change what it is used for. How will most laws treat you?

  3. 3.Why does the Academy assessment cap your level at 'Emerging' when inventory or risk classification scores are low?

0 of 3 answered

Sources

Last reviewed Sep 25, 2026.

Educational information, not legal advice. Laws change and details depend on your situation — check the linked sources and talk to a qualified lawyer before acting. Last content review: 2026-09-25.

Spotted an error? Ask the tutor or email hello@myaiguide.pro.