Published on Oct 08, 2026
Artificial intelligence is becoming increasingly embedded in products, services and business processes. For commercial lawyers and contract managers, that creates a practical question:
“Do existing contracts adequately address the risks and responsibilities that come with AI?”
Traditional provisions covering intellectual property, confidentiality, data protection, warranties, liability and service levels remain important, but AI can introduce additional questions that may not have been contemplated when the contract was drafted.
The challenge is not necessarily to create an entirely new AI contract. In many cases, it is about identifying where existing contractual protections need to be adapted.
One of the first challenges is surprisingly basic: what exactly does the contract mean by AI?
AI can encompass a wide range of technologies and applications, from generative AI tools to AI-enabled software and automated decision-making systems. A broad or unclear definition could create uncertainty about which contractual obligations apply.
Contracts may therefore need to establish what types of AI systems, tools or uses are within scope and whether particular obligations apply to the supplier, customer or both.
It is also important to understand how AI will actually be used under the agreement.
· Will a supplier use AI to provide the contracted service?
· Will customer data be processed by an AI system?
· Can a supplier use information supplied by the customer to train or improve its models?
These questions can have significant implications for confidentiality, data protection and intellectual property.
Contractual restrictions around the use of customer data for AI training may therefore be particularly important where commercially sensitive or personal information is involved.
Intellectual property (IP) is another area where AI can complicate familiar contractual concepts.
Parties may need to consider who owns or has rights to use AI-generated outputs, what happens where third-party material has been incorporated into an output, and whether the supplier can provide the customer with the rights it expects to receive.
The contract should also distinguish between the customer's existing materials, supplier technology, AI models, training data and newly generated outputs where these are relevant to the relationship.
AI systems can produce inaccurate or inappropriate outputs. The contractual question is therefore not simply whether an AI system works, but what happens when it does not.
Depending on the service, parties may need to consider warranties, accuracy obligations, service levels, indemnities and limitations of liability.
There may also be questions around responsibility for reviewing AI-generated outputs. A contract that allocates responsibility clearly can help reduce uncertainty when an AI-enabled service produces an unexpected result.
AI technology and its use are developing rapidly. A contract negotiated today may need to remain workable as systems, business practices and regulatory expectations evolve.
This makes mechanisms such as change control, review provisions, audit rights, termination rights and appropriate information obligations potentially important.
The objective is not to predict every future development. It is to create a contractual framework that gives the parties sufficient clarity while allowing the relationship to adapt.
The most effective approach to AI contracting is unlikely to be a collection of generic ‘AI clauses’ inserted into every agreement.
The appropriate provisions will depend on the technology, the services being provided, the data involved, the parties' respective roles and the risks associated with the particular use case.
For commercial lawyers and contract managers, understanding those underlying issues is therefore becoming increasingly important. The starting point should be the commercial relationship and the risks it creates, rather than the technology alone.
IPI Academy's Commercial & Business Law training course portfolio includes several courses covering the legal and contractual issues surrounding AI and technology, including:
Published on Oct 08, 2026 by Angela Spall