Opinion: Simon Martin
In the last couple of years, Generative AI has made substantial strides in the business sector, with the potential to fundamentally transform processes within organisations. The adoption of AI technologies is being encouraged across various industries globally, as businesses recognise the benefits of integrating these tools into their daily practices. The key advantages include streamlining operations to enhance efficiency, reducing reliance on human intervention for repetitive tasks and leveraging existing data to improve production processes and service delivery.
From startups to established enterprises, there is a push to harness the capabilities of Generative AI – the advantages of doing so are becoming increasingly evident. Organisations that embrace these technologies position themselves for significant improvements in performance and customer satisfaction.
By automating repetitive tasks and generating content or solutions that were once time-consuming, businesses can reallocate valuable resources to more strategic initiatives. However, like any innovation, this advancement also presents certain risks (both legal and commercial) that must be effectively managed within a business context. For every organisation poised to embrace this emerging technology, comprehending and navigating these risks is the first and most critical step in their AI journey.
Implementing Generative AI within your business—along with determining which AI to integrate—constitutes a significant business decision. It is important to incorporate the legal considerations that accompany Generative AI into this decision-making process.
We recommend that the following three key issues be taken into account:
Looking at each of these issues in a little detail:
Intellectual Property Considerations in AI Licensing
When dealing with intellectual property (IP) in the context of AI licensing, it is crucial to understand whether there is any risk that the company licensing the AI may not possess the rights to the materials on which the AI tool has been trained. Currently, several legal cases are underway, addressing whether companies that have developed AI tools were permitted to utilise public materials (such as materials on the internet) to create their algorithms and train the tools they now market.
Major companies engaged in AI are increasingly aware of these concerns, and their terms of service typically clarify what steps they will take should they lack the necessary rights. Some companies are even providing their customers with generative AI indemnification if those customers are challenged on copyright grounds. When evaluating whether to implement an AI tool—and determining which specific tool to choose—it is wise to scrutinise the vendor’s statements (and their terms of supply) regarding
IP rights.
Should you decide to adopt an AI tool, you should also pay particular attention to the terms regarding ownership of the output generated by the AI. While it is becoming less common, some terms still assert that the vendor retains ownership of the output, providing you with a licence. If you intend to build a significant asset of your business around the AIgenerated output, you need to ensure that you possess the requisite rights.
Lastly, consider whether the manner in which your AI-assisted process generates output allows for IP rights to be attached to that output. Notably, the laws of some countries presuppose the involvement of a human in the creative process. Assess whether your process includes sufficient human intervention to establish ownership of the output and ensure that it qualifies as IP.
Data and Privacy Considerations
The issue of data and privacy hinges on the terms delineated by your vendor regarding the handling of your data. Generative AI models are developed utilising vast amounts of accessible data, and their value is rooted in their ability to refine approaches based on widely accepted norms. The accuracy of the methods adopted by these models is directly proportional to the volume of data they can access. Consequently, vendors have a vested interest in acquiring more data.
It is important to ensure that you understand the basis upon which the data you provide will be utilised by the vendor. Should the vendor employ your data not solely to deliver a service but also to enhance its tools, you must take into account the following considerations:
Changes
Once you have successfully navigated the potential Intellectual Property (IP) and Data issues, and have established a comfort level regarding these matters, evaluate whether the vendor retains the right to amend the terms at any time.
It is understandable that vendors require the flexibility to alter the terms under which they provide services—especially when catering to thousands of customers. This model permits a more economical service through a one-to-many approach. The economics of the model assume that everyone is on largely the same terms. As such this model requires the ability to modify these terms unilaterally. If to change the terms the vendor had secure agreement from all customers this would render this business model unfeasible.
As a business poised to implement an AI tool, you will need to assess the risk associated with potential changes to the terms that could inhibit your ability to continue utilising the service. Often the vendor offers you the right to discontinue use of the service, alongside ceasing payment obligations. Should this option be available, make sure that you do not incur charges for any period in which the tool is not utilised and that you have the right to receive a refund for any prepaid amounts corresponding
to unused time.
However, it is equally important to evaluate the practicality of discontinuing use. Once the tool is in operation, consider how seamless it will be for you to cease its use. This evaluation should be an integral part of implementing the AI tool—determine whether you can establish a straightforward exit strategy. The ease of doing this might be better with some solutions rather than others – if that is the case then this should be taken into account as part of your choice of AI tool.
The likelihood is that more and more AI tools will become available and some will be inexpensive (or free) and easily accessible. To avoid well intentioned members of your team signing up for AI tools or using AI
tools in their day to day work without ensuring that the above issues are squared away it would be a good idea to have a clear statement on how these issues should be considered and addressed.
As a firm we did this with our AI policy. You should carefully consider what your policy would be and how you make that known to your team.
For more information on ensuring your AI strategy is legally sound, download our free practical guide, which provides insights and actionable steps to help you navigate this transformative journey effectively.
