CLS Blue Sky Blog

Sullivan & Cromwell Discusses Use of Artificial Intelligence Tools in M&A Transactions

Summary

Use of AI Tools in M&A Transactions

Machine learning AI tools have been used in M&A deals for several years. The primary focus of these tools has been to increase efficiency during due diligence review by, for example, identifying and extracting specific contractual provisions from documents in a virtual data room (“VDR”). However, since the advent of large language models, GenAI has broadened the potential applications of AI tools to legal practice, including on M&A deals. M&A attorneys have been grappling with how to incorporate emerging legal GenAI technology in their workflows in a manner that increases efficiency and improves attorney work product, without sacrificing accuracy or otherwise undermining client confidence, legal judgment, or compliance with professional obligations.

Emerging Uses of AI in the M&A Process

The use of AI tools is emerging in many stages of the M&A process, including for market and legal research, during the due diligence process, in negotiating and drafting transaction documents, and for deal execution. Emerging use cases for AI tools during the M&A process include:

Research

Due Diligence

Transaction Documentation Drafting and Negotiation

Deal Execution

Risks of Using AI Tools in M&A practice

Despite the promise of AI, M&A attorneys leveraging GenAI tools should be aware of the various risks presented by AI tools. GenAI tools are prone to making surprising errors and “hallucinations,” which may be phrased in a confident tone but are nonetheless false, misleading, or incomplete. There have been numerous instances of attorneys filing legal documents containing well-formatted case citations to nonexistent cases. AI tools may also miss important clauses or overlook hidden liabilities.

Practitioners should also consider the assurances clients and other stakeholders (e.g., representation and warranty insurers) may seek with respect to the accuracy and completeness of attorney work product (e.g., diligence summaries and reports).

GenAI products may also exhibit biases resulting from the data on which they were trained—for example, a bias in favor of publicly filed transactions, which are more readily available and therefore more commonly included in AI training datasets.

Although AI tools are improving in these regards, we still routinely observe inaccurate, incomplete, misleading or biased output from GenAI tools. Accordingly, all work produced by GenAI should be carefully reviewed by qualified lawyers.

It is also important for attorneys to understand how data provided to AI tools is used by the tools’ developers. Unless appropriate data boundaries (both legal and technical) are put in place, lawyers risk breaching obligations of confidentiality or inadvertently waiving privilege by disclosing confidential or privileged materials to AI tools. Attorneys should also consider both the accuracy limitations of GenAI and applicable regulatory or documentation retention requirements before enabling AI features that automatically summarize calls and other communications.

Additionally, legal practitioners should ensure that their use of AI tools comports with their ethical obligations, including the duties of competence and confidentiality, and remain abreast of any updated rules or guidance that directly address the use of AI for client matters.

Conclusion

AI tools, and particularly GenAI, are beginning to transform the M&A process by streamlining research, due diligence, drafting, and deal execution. Yet these benefits come with risks, including accuracy errors, biases, confidentiality concerns, and ethical obligations. Attorneys should adopt these tools thoughtfully, with rigorous human oversight and clear internal policies.

This post is based on a Sullivan & Cromwell LLP memorandum, “Use of Artificial Intelligence Tools in M&A Transactions,” dated January 12, 2026, and available here. 

Exit mobile version