Artificial Intelligence (AI) tools are increasingly shaping how organisations operate. Whilst the benefits of adopting AI can be significant, it is crucial to understand the terms you are agreeing to before signing up with a provider.
AI software contracts (often in the form of terms and conditions of service) are typically drafted in favour of the provider. If not carefully reviewed, they may expose organisations to unnecessary risks. Below are five key terms that organisations should pay particular attention to in AI software contracts.
Limitation of Liability
Most providers will limit or cap their liability to users in various ways, for example by:
- Capping their liability to the fees paid in the previous 12 months. Organisations should consider whether such a sum would cover its potential losses.
- Excluding liability for any indirect or consequential losses.
- Excluding liability for specific breaches, such as confidentiality, data protection or intellectual property rights.
This means that if the AI produces incorrect or infringing outputs that causes your organisation a significant loss or damage, your ability to recover losses may be significantly restricted.
Warranty and Indemnity
Providers will generally give minimal warranties and indemnities to organisations. Common examples include:
- The software functioning only “substantially” as described.
- No assurance that outputs will be accurate, error-free and fit for your business purposes.
- Disclaim any responsibility for service interruptions, delays, or delivery failures.
Again, this means that if the AI falls short of your organisation’s operational needs, you may have limited recourse against the provider.
Insurance
Whilst some providers will carry insurance (e.g. professional indemnity etc.), organisations should consider whether:
- The provider is contractually obliged to maintain appropriate insurance cover for the duration of the contract.
- The level of insurance cover and the type(s) of insurance cover is sufficient to meet your potential losses (subject to any liability cap as discussed above).
- Evidence of insurance cover can be obtained on your request, or at agreed intervals.
Without the above, recovery in the event of a loss or damage could be uncertain.
Intellectual Property and Data Rights
Ownership and use of data and outputs are significant issues when adopting AI. Key risks include:
- Providers retaining ownership of your data and any outputs.
- No guarantee that outputs will not infringe the intellectual property rights of any third-party.
- Restrictions on your organisation’s ability to use the outputs generated.
- Inadequate obligations on the provider to assist your organisation with your compliance under data protection law.
Organisations should ensure that any contractual provisions in this respect align with your business needs, particularly where data use is critical.
Termination and Service Changes
Termination provisions and service changes are often weighted heavily in the provider’s favour, for example:
- Short notice periods for termination by the provider, and under a wide range of circumstances compared to those available to organisations as the customer.
- Providers reserving the right to unilaterally amend features, pricing, or terms.
- No obligation to return or confidentially destroy your data on termination or expiry of the contract.
The above provisions can significantly affect your business continuity and data protection compliance.
Conclusion
AI software can be helpful to an organisation’s operations, but the legal terms governing its use are just as important as the technology itself. Terms and conditions should be reviewed carefully – ideally with legal advice – to ensure that they provide appropriate protections and align with your organisation’s business needs.
If you, or anyone in your organisation, requires further advice or assistance in this regard, please contact our team.
