Artificial Intelligence (AI) is no longer just a buzzword. From automated customer service to data-driven decision-making, businesses across all sectors are increasingly integrating AI into their operations. But while AI offers efficiency, scalability and innovation, it also brings new legal questions to the table, especially when it comes to AI in commercial contracts.
As businesses enter into supplier, licensing, outsourcing or joint venture agreements involving AI technologies, it’s essential that your contracts reflect the realities and risks associated with this rapidly evolving technology. In this article, we look at the seven key issues businesses should consider capturing within their AI Contracts:
1. Clearly Define AI Roles and Functions
Before negotiating terms, it’s critical to be clear about what the AI system is expected to do. Contracts should include a detailed description of:
- The specific AI functionality or service being provided
- Whether the AI is standalone or part of a wider system
- If machine learning is involved, how and when the system will be trained or updated
Ambiguity around AI functionality can lead to disputes over delivery standards, performance failures or system integration issues.
2. Allocation of Risk and Liability
When AI is involved in generating outputs, making decisions, or processing personal data, the question of liability becomes more complex.
Key considerations:
- Who is responsible if the AI system malfunctions or produces incorrect results?
- What happens if those results cause financial loss, reputational damage or regulatory breaches?
- Can traditional warranties and indemnities still be relied on?
You may need to update or supplement your standard clauses to allocate risks appropriately. This could include, mutual indemnities, AI-specific warranties, and tailoring your limitation of liability to the AI use case
3. Intellectual Property Ownership
With generative AI now widely adopted, businesses must be clear on who owns the outputs.
Your contract should answer:
- Who owns the data used to train the AI?
- Who owns the content or product generated by the AI?
- Is the AI tool itself proprietary or open source and are there any there licensing restrictions attached to its use?
If you’re the client or licensee, be cautious of terms that assign IP in the AI-generated content to the provider. You may need to negotiate exclusive or joint ownership depending on the nature of your business and intended use of the output generated.
4. Confidential Information and Datasets
Before uploading data to any AI tool it’s vital that the content and nature of the datasets are carefully considered to ensure intellectual property rights and other confidential information of the business and/or its customers isn’t being used without the appropriate consents or safeguards in place, including;
- Robust confidentiality obligations
- Usage parameters are established, including whether datasets can be used to train the AI
5. Data Protection and Regulatory Compliance
AI systems often rely on vast amounts of data to function effectively. Where personal data is involved, your commercial contract must address compliance with UK GDPR and other regulatory requirements.
You should ensure that the contract covers:
- Data controller vs data processor roles
- Lawful basis for processing personal data
- Data security measures and technical safeguards
- Audit rights and cooperation duties
The changes being phased in by the Data Use and Access Act 2025 (DUAA) are an example of why it is important to ensure your contracts keep pace.
6. Performance Standards and Service Levels
Traditional service level agreements (SLAs) are unlikely to be an appropriate method for measuring the performance of AI-based services, particularly those that involve probabilistic outputs (e.g. chatbots, fraud detection tools, recommendation engines) as they really cater for data accuracy or quality.
Consider how to:
- Define acceptable accuracy and quality thresholds
- Monitor AI performance over time
- Include re-training or improvement obligations
As AI and related technology are evolving rapidly, you should consider including flexibility in the service level parameters and a process for agreeing periodic updates.
7. Ethical and Responsible AI Use
Whilst there is currently no specific AI legislation in the UK, we are awaiting an AI Bill. This however does not mean that AI can be operated without regard to existing legislation (including the UK GDPR) or in an unethical manner and businesses should proceed with caution as those operating AI or making use of decisions based on AI can be held responsible should things go wrong, such as biased or discriminatory decisions. It’s therefore worth including clauses relating to ethical AI usage. This might include:
- Ensuring the AI is not used to discriminate unlawfully
- Ensuring AI is trained on diverse datasets and audits obligations
- Ensuring there is human intervention or oversight within the AI process
- Prohibiting use for surveillance or profiling without explicit agreement
- Requiring suppliers to comply with industry standards or codes of practice
This area is still emerging in commercial contracting but can help mitigate reputational risk and align with ESG goals.
8. Exit and Transition Planning
Finally, your contract should address what happens if at the end of the contractual relationship.
Particularly with AI in commercial contracts, it’s important to plan for:
- Access to data and outputs post-termination
- Smooth handover or transition to a new provider
- Destruction or return of proprietary training data or system logs
These steps can be crucial in protecting your IP, trade secrets and business continuity.
How Slater Heelis Can Help
As AI becomes more embedded in business functions, commercial contracts must evolve to match. Off-the-shelf templates and outdated clauses simply won’t do the job.
At Slater Heelis, our Commercial team works closely with clients to tailor agreements that reflect both the opportunities and the risks of AI. Whether you’re a provider of AI solutions, a business integrating AI into your services, or simply exploring your options, we can help you contract with confidence.
Get In Touch
Carla Murray is a Partner and the Head of our Commercial team, she specialises in technology law, data protection, commercial contracts, and IP.
Need help reviewing or drafting an AI-related contract? Get in touch with our team today to book an initial consultation by filling out our online contact form or call 03300 297 347.
