Heyl Royster


Heyl Royster


Using AI Technology to Extract/Analyze Information in Legal Contracts


By: Michael Kokal, mkokal@heylroyster.com

Law firms and corporate legal departments face challenges arising from the sheer number of contracts they must keep track of, organize, and update. Many firms don’t have a database of all the information in their contracts, not to mention an efficient way to extract and review relevant data. However, new AI technology, which uses a combination of natural language processing (NLP) and machine learning, offer potentially new and exciting tools to extract and analyze data quickly.

The possible applications of this technology are vast. Take the case of the ubiquitous non-disclosure agreement (NDA). For most business people, the NDA is very much a standard document. However, there are pitfalls in the fine print that many people don’t take the time to read. Moreover, the business person might not have time to “go back to legal” to have an NDA reviewed if time is of the essence for completing a deal. In this scenario, companies are developing programs that use machine learning and NLP to assist the business person in quickly reviewing a contract. These new software protocols will ultimately inform the business person that the document is “ok to sign,” or identify the areas that need to be re-negotiated, or whether ultimately the document needs to be reviewed by a lawyer because of draconian terms.

As an example, the non-disclosure agreement will contain a certain “term” length. However, there may be a standard “term” that is the custom and practice in the industry or in your geographic area. If the non-disclosure agreement contains a term in excess of this standard, the AI software should quickly identify and flag the provision and advise the business person not to sign, and further advise them of “customs and practices” in the industry assuming these can be identified by the programmer.

Another example of the NDA in the tech space is the “residuals clause.” The clause, in some ways, can be tantamount to a “back-door IP license.” Software programs could identify the existence of a residuals clause in an agreement and inform the business person that they either have to remove it, which might be a common suggestion, or, perhaps, sign the clause, but with restrictions. In any event, it could put both of the parties on notice as to the existence of the clause in order that an agreement on the scope could be quickly negotiated.

Further examples are endless. Indemnification clauses could be analyzed to provide notification when a certain cap is reached. Insurance agreements could also be pre-programmed. What is clear is that AI in the legal space will only be increasing in the future. Savvy attorneys who know how to draft an AI contract for an AI audience would distinguish themselves from their competition and demonstrate value to their clients, not only in their legal acumen, but in their ability to anticipate client’s business needs and interfacing with these various programs.