As artificial intelligence ("AI") becomes more topical each day, more organisations are wanting to incorporate AI tools in their business operations, or use the AI tools to enhance the organisations' existing service offerings to end users. We highlight the three main procurement models to choose from when adopting an AI tool for business purposes.
Before procuring and deploying any AI tools, you should determine your organisation's risk appetite. In doing so, you might consider:
- how the AI service will be integrated into enterprise systems and service offerings;
- how critical the AI functionality will be to the organisation or its service offerings;
- the degree of commercial, data, and security risks which the AI system will expose the organisation to; and
- which risks the organisation would be willing to underwrite in the event that the AI service provider is not amenable to negotiating its standard terms and conditions in respect of the organisation's procurement or use of the AI tool.
Next, you will need to determine which, out of the following three AI procurement models (or a combination thereof), is best suited for the organisation's business needs and requirements:
- employing resources who are solely dedicated to the development of AI tools specifically and exclusively for the company, ie, in-house developers
One of the main roles and responsibilities of in-house developers would be to develop AI tools that the company wishes to incorporate into its business operations, or otherwise to develop customisations or enhancements to the company's existing service offerings to include AI functionality. This development would be done in the course and scope of employment, and accordingly ownership in any developments created by such in-house developers will be governed by their employment contracts. The employment contracts should therefore contain intellectual property (specifically copyright) assignment clauses to ensure that the intellectual property (including moral rights) in any tools or systems developed by such personnel pursuant to their role within the organisation is automatically assigned to the company.
- outsourcing the development of AI tools to third party developers, ie, external developers
When procuring external developers to create and develop AI tools specifically for the company, it is important that the company concludes a development agreement with the third party. This agreement should contain favourable clauses to the company, and require the third party to waive any rights over the source and object code in all developed intellectual property, as well as include provisions that give effect to the automatic assignment of all intellectual property rights (and any moral rights) in developed intellectual property immediately upon the creation or development thereof. To the extent that the developer makes use of any third-party intellectual property in the development of the AI tool for the company, the company should request appropriate indemnities to protect itself against any third party claims in respect of intellectual property infringement or unauthorised use of the third-party's intellectual property.
- procuring and deploying a third-party AI tool which is available to the public at large
The procurement of developed AI tools from a third-party is convenient as it eliminates the costs associated with the company having to develop its own AI tools. However, these AI tools will not be bespoke to the company and will not provide the company with competitive advantage in so far as AI functionality is concerned. Since the AI tool belongs to the third party, the company will merely be granted a licence to use the AI tool for specific purposes, and will therefore not assume ownership or otherwise have control over the AI tool itself. Accordingly, it is important for the company to carefully review and negotiate the third party's terms and conditions or end-user licence agreement to reduce its risk exposure, and to determine the risks that the company is willing to underwrite where the third party AI provider is unwilling to negotiate standardised terms.
The main clauses that the company should seek to negotiate relate to, inter alia, liability, intellectual property, warranties (regarding the functionality and performance of the AI tool, fitness for a particular purposes, etc.), and indemnities (in respect of unauthorised use, third party infringement claims, data breaches, cyberattacks, and the like) which will reduce the company's risk exposure in respect of the use of the AI tool, its decision-making, or the outputs generated by such tool, thereby protecting the company from, and limiting its liability against, unnecessary, unexpected, and excessive loss or damages.
Furthermore, the agreement should clearly stipulate the licensing terms in respect of the company's use of the AI tool and provide the company with the right to use, copy, modify, and integrate the AI tool into its organisation for internal and/or business purposes.
Once the company has determined the AI model that is best suited for its organisational needs and requirements, it is recommended from a governance perspective that the company completes an ethical impact assessment, establishes an AI ethical review committee, implements measures for the continuous review of the performance and functionality of the AI tool, and implements business continuity or fault tolerance measures to protect the company in the event that the AI tool's fails or experiences unplanned or excessive downtime.
Furthermore, the company should also implement an AI procurement plan which deals with the following aspects:
- business and commercials objectives;
- data regulation (including removing data bias); and
- governance and compliance.
The procurement and use of AI tools may introduce great commercial benefits for companies; but companies should take steps to reduce their risk exposure in order to reap the benefits by conducting a thorough risk analysis, reviewing and negotiating the vendor terms and conditions, and developing an internal AI procurement strategy. The implementation of a sound AI procurement framework is essential for increasing the maximum return on investment derived from the company's procurement and use of AI tools.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
