Artificial Intelligence in M&A: Efficiency, Due Diligence and Risk Management

AI has been an important and increasingly prevalent feature of M&A transactions for some time. Buyers need to understand how AI is used in the target business (to understand and evaluate potential risks arising from that use and to plan for how the AI assets of the target can best be integrated with those of the buyer) and businesses that are identified as part of the “AI sector” itself are increasingly the subject of heightened levels of M&A activity themselves. In addition, buyers and sellers are increasingly using AI systems in various aspects of M&A transactions, irrespective of the types of businesses being bought and sold.

How is AI transforming M&A due diligence and deal implementation?

M&A lawyers have in recent years been deploying AI tools to process large amounts of due diligence (DD) data in order to identify issues that are the focus of their DD processes. There are a number of commercially available products that will read the entire contents of a data room and more or less immediately identify features which would otherwise take individuals significant amounts of time to find. Until recently it has generally been machine learning that has been used in this kind of review, but GenAI solutions are quickly emerging in this area as well.

The technology can also quickly identify clusters of documents that are in a similar form, for example, they might all be based on a template distribution agreement. The technology can equally quickly highlight variances, for example, documents which are missing wording (or even whole clauses), documents which have additional clauses and documents where template clauses have been varied.

Similarly, the technology can screen a data room to pull out all of the documents which contain specific types of clause, such as a change of control clause. This requires the technology to utilise learning that can do this (as a change of control clause does not need to contain the word “change” or “control”). The model is trained on document sets that have gone before and if it is in doubt it checks with the human user – based on the response it is given it “learns” by refining itself, having taken into account the new information.

AI solutions can also assist in the preparation of draft deal documentation and to overlay existing machine learning tools in the comparison of draft documents (received from the other side of the transaction or produced internally) with a specific data set. This could for example be the internal library of similar documents held by a law firm or legal department, or an external data set such as documents filed with the SEC in the US and stored on its Edgar system. In addition, there is growing use of AI tools that summarise anything from the contents of diligence materials to the reasonableness of the negotiating position taken in a draft deal document, as well as tools that assess market terms taken from large numbers of sources and highlight current market positions and trends. Clearly defined policies and governance will be required in order to deploy all this technology and explain to clients the benefits and risks involved.

How can AI-related risks be identified, managed  and mitigated in M&A transactions?

A key component of any DD review going forward will be the assessment of AI-related risks in a target. For example, are the target or its personnel deploying AI in a way that breaches the law or the rights of third parties? There are many examples  of how such violations might occur and of course the particular issues will often depend on the type of the underlying business. Consider the following:

• Is AI being used in a business in a way that is potentially discriminatory (for example in making decisions between job candidates or, in financial services businesses, when deciding whether to give a person credit)?
• Are AI products being used by target personnel in a way which might compromise the confidentiality of confidential information, or which might enable AI platforms to use that data in potentially unexpected ways?
• Is AI being used in a way that unfairly processes personal information in breach of data privacy requirements?
• Do the issues around the potential inaccuracies produced by GenAI (“hallucinations” etc) create risks for the business?
• Does the use of AI in the business risk infringing the IP rights of third parties, or of itself creating IP which is not sufficiently protected?

With the legal landscape in a state of flux as different jurisdictions find alternative ways to regulate the growing use of AI in businesses, these issues are becoming a fast-moving target where buyers in particular need to lean on the expertise of their deal lawyers.

How is AI driving M&A activity within the technology sector itself?

As the use of AI grows rapidly, businesses which have the AI technology and/or the talent powering that growth have become increasingly in demand as acquisition targets themselves, with smaller AI companies often getting acquired before they can grow significantly.

Because of the speed of development of AI, acquiring an AI startup is often a more efficient way to integrate new AI tools and solutions into existing products and services and some of the major tech players have been particularly active in AI acquisitions.
Specialist AI talent is also often a key driver in such acquisitions, allowing larger businesses to accelerate their AI development without having to build expertise from scratch.

What’s next for AI and M&A

The AI-related risks referred to above and the fast-changing legal and regulatory developments governing them are likely to   have a significant impact in relation to  AI businesses, given the scope of such changes to enhance, diminish or even destroy their value.