---
title: "AI Contracts | Pitch, European law firm for IP, IT & data"
description: "Contracts for buying, licensing and building AI systems, drafted or reviewed with the AI Act roles in mind and clear terms on data, intellectual property,…"
url: https://pitch.law/solutions/ai-contracts
---

[pitch](https://pitch.law/) [Solutions](https://pitch.law/solutions) [AI](https://pitch.law/solution-categories/ai)

# AI Contracts

Contracts for buying, licensing and building AI systems, drafted or reviewed with the AI Act roles in mind and clear terms on data, intellectual property, performance and liability.

- - By [Bart Lieben](https://pitch.law/team/bart-lieben), Attorney-at-Law
  - 19 November 2026

- Updated

  19-11-2026

## The question

You are about to buy an AI tool, license a model, or have a developer build an AI system for you. Or you are the one selling it. The supplier's standard terms say little about what happens to your data, who owns what the system produces, what it must achieve, or who carries the AI Act duties. What should the contract say?

The AI Act (Regulation (EU) 2024/1689) gives the starting point. A provider is the person who develops an AI system, or has it developed, and places it on the market or puts it into service under its own name or trademark (Article 3(3)). A deployer is the person using an AI system under its authority, outside personal use (Article 3(4)). The duties differ a lot between the two ([provider or deployer](https://pitch.law/academy/provider-vs-deployer-ai-act)), and a contract cannot move them: a company that commissions a system and launches it under its own brand is the provider, whatever the contract calls it. A deployer, distributor or other party can also become the provider of a high-risk system, for instance by putting its name or trademark on it (Article 25(1)).

The Act also reaches into contracts directly. Where a third party supplies an AI system, model, tool, service or component used in a high-risk system, the provider and that supplier must specify by written agreement the information, capabilities, technical access and other assistance the provider needs to comply (Article 25(4), as amended in 2026). A deployer of a high-risk system must use it according to the instructions for use (Article 26(1)), so you need those instructions, and the right to get updates. And if the system is built on a general-purpose AI model, its provider must have a copyright policy and publish a summary of the content used for training (Article 53(1)(c) and (d)).

## Our approach

We start with the facts. What does the system do, who built it, where does it run, which data goes in, and under whose name does it go to market? From that we settle the AI Act role of each party and the risk class of the system. Without that answer the rest of the contract rests on guesswork.

Then we go through the points that matter in AI deals:

- **Data.** What you supply, whether the supplier may use it to train or improve its models, how long it is kept, and how it is returned or deleted. Personal data needs a processing agreement as well ([data processing agreements](https://pitch.law/solutions/data-processing-agreements)). Licensed training data has its own terms ([data for AI](https://pitch.law/academy/data-for-ai-data-spaces-and-licensing-training-data)).
- **Intellectual property.** Who owns the model, the fine-tuning, the prompts and the output, and what licence each side gets. See [AI and intellectual property](https://pitch.law/solutions/ai-and-intellectual-property).
- **Performance.** What the system must do, how that is measured, what happens when accuracy drops or the model is changed, and who may audit.
- **AI Act duties.** Who does what: documentation, instructions for use, logging, human oversight, transparency to users, incident reporting, cooperation with authorities.
- **Liability.** Caps and indemnities that follow the risk each party can control, including third-party claims about training data or output.
- **Exit.** Access to data and logs at the end, and a way out if the system no longer complies.

For a supplier's draft we use [our contract review](https://pitch.law/technologies/clause-library-risk-review): each clause is read against our playbook and scored for risk, so you see where to negotiate first. When we draft, [our Document Studio](https://pitch.law/technologies/contract-studio) assembles the contract from our clause library in a fixed structure and delivers a Word file. Both run on our own AI, and a lawyer checks every finding and every clause before it reaches you.

In public procurement, the model contractual clauses for AI published on the European Commission's Public Buyers Community (updated in March 2025, in a version for high-risk AI and a lighter one) are a useful reference for public buyers, and we use them as a checklist where they fit.

For more background, read [AI clauses in vendor contracts](https://pitch.law/academy/ai-clauses-in-vendor-contracts) and our article on [codes, guidelines and standards for general-purpose AI](https://pitch.law/academy/general-purpose-ai-codes-guidelines-standards).

## What you get

- A note on the AI Act role of each party and the risk class of the system.
- For a supplier's terms: a clause-by-clause review with a risk ranking and proposed redlines.
- For your own deals: a contract or AI addendum drafted from our clause library, in English, Dutch, French or German.
- Where needed, a processing agreement and a data licence that match the AI contract.
- Our negotiation with the other side, if you want it.

Limits apply. A contract allocates duties between the parties; it cannot change who the AI Act treats as provider or deployer, and it cannot make a prohibited use lawful. A warranty is only as good as the party giving it, so we also tell you which information to ask for before you sign.

## Next step

Send us the supplier's terms or your draft, with a short description of the system and how you will use it.

[Schedule a meeting](https://pitch.law/schedule) or [contact us](https://pitch.law/contact-us).

## In short

- Your AI Act role checked before the contract is signed
- Clear terms on input data, training and output
- Ownership and licences of models and results settled
- Performance, oversight and incident duties made workable
- Liability allocated to the risk each party controls
- Reviewed by a lawyer

## Our proprietary Legal Tech

- [Technology](https://pitch.law/technologies/contract-studio)
  ### [Document Studio](https://pitch.law/technologies/contract-studio)
  [Our Document Studio drafts and analyses legal documents in a secure, on-premise environment, and keeps the documents behind your matters up to date as new data comes in and procedural steps fall due.](https://pitch.law/technologies/contract-studio)
- [Technology](https://pitch.law/technologies/clause-library-risk-review)
  ### [Contract Review](https://pitch.law/technologies/clause-library-risk-review)
  [Send us the contract the other side drafted and we check it clause by clause against our playbook, with a risk ranking an attorney has verified.](https://pitch.law/technologies/clause-library-risk-review)

## Background reading

- [Academy](https://pitch.law/academy/ai-clauses-in-vendor-contracts)
  ### [AI Clauses in Vendor Contracts: 9 Terms to Negotiate](https://pitch.law/academy/ai-clauses-in-vendor-contracts)
  [The AI clauses to negotiate when you buy or supply AI: AI Act roles, training on your data, output ownership, IP indemnity, audit and logs, liability.](https://pitch.law/academy/ai-clauses-in-vendor-contracts)
- [Academy](https://pitch.law/academy/provider-vs-deployer-ai-act)
  ### [AI Act Provider vs Deployer: Duties and Article 25](https://pitch.law/academy/provider-vs-deployer-ai-act)
  [Under the AI Act, the provider builds or brands an AI system and the deployer uses it. Each role's duties, and when Article 25 makes a deployer a provider.](https://pitch.law/academy/provider-vs-deployer-ai-act)
- [Academy](https://pitch.law/academy/data-for-ai-data-spaces-and-licensing-training-data)
  ### [Data for AI: data spaces and licensing data for training](https://pitch.law/academy/data-for-ai-data-spaces-and-licensing-training-data)
  [Where AI developers can lawfully get training data in the EU, what the law says about common European data spaces, and how the Data Act, the AI Act's data rules and the text and data mining exceptions shape a training-data licence.](https://pitch.law/academy/data-for-ai-data-spaces-and-licensing-training-data)
- [Academy](https://pitch.law/academy/general-purpose-ai-codes-guidelines-standards)
  ### [General-Purpose AI: Codes of Practice, Guidelines and Standards](https://pitch.law/academy/general-purpose-ai-codes-guidelines-standards)
  [What the AI Act requires of providers of general-purpose AI models in Articles 51 to 56, and how much weight the code of practice, Commission guidelines, harmonised standards and codes of conduct carry.](https://pitch.law/academy/general-purpose-ai-codes-guidelines-standards)

## More in AI

- ### [AI Act Compliance](https://pitch.law/solutions/ai-act-compliance)
  [We help you find out which of your AI systems fall under the EU AI Act, in what role, and what you must have in place and by when.](https://pitch.law/solutions/ai-act-compliance)
- ### [AI Governance and Policies](https://pitch.law/solutions/ai-governance-and-policies)
  [An AI policy your teams can follow, with an inventory of the tools you use, clear roles, an approval route for new tools and rules for acceptable use, linked to your GDPR work.](https://pitch.law/solutions/ai-governance-and-policies)
- ### [AI Literacy Training](https://pitch.law/solutions/ai-literacy-training)
  [Training sessions on AI and the law for your staff, matched to the tools they use and their role, with materials and an attendance record you can keep as proof of the measures you took.](https://pitch.law/solutions/ai-literacy-training)
- ### [AI and Intellectual Property](https://pitch.law/solutions/ai-and-intellectual-property)
  [Advice on the intellectual property questions AI raises, from training data and text and data mining to who owns what a system produces and how you keep your own know-how out of other people's models.](https://pitch.law/solutions/ai-and-intellectual-property)

## Let’s talk

Tell us what you’re working on. We’ll get the right person on it and reply within one business day. The first chat is free.

[Schedule a meeting](https://pitch.law/schedule)
