What are the benefits and risks of using AI tools for contract negotiation?

This scenario is great for a w2 type of employment with an organization that fosters long term employment and professional growth. The ideal scenario is to have leaders and managers that have strong communication skills and thrive on having crucial conversations where difficult discussions may need to be had. This is typically not the case when your employment is on a contractual basis and often times the end client is unwilling to provide direct performance feedback as it may conflict with fair and open contract solicitations.

Relationship-based negotiations prioritize the human aspect of business interactions. They recognize that successful negotiations are not just about reaching an agreement but also about fostering connections that endure beyond the negotiation table.

Relationship-based negotiations also have a cultural aspect to them. Many cultures in the world require a baseline of trust and foundational relationship to effectively do business and negotiate new deals. These cultures often require significant personal investment upfront and demonstration of honesty and integrity prior to the nitty-gritty of contract terms. But, the dividends are that the business relationships can net long-term and reliable partnerships if managed properly.

This is a typical arrangement in a staff augmentation or contractual relationship, where the relationship aspect is minimized in lieu of streamlined contracting practices. This has its advantages as well if navigated correctly, but isn’t ideal for a direct w2 relationship where the goal is a long term career with personal growth aspirations within the organization.

Transactional negotiation does not need the same level of personal investment as relationship building, but trust and transparency can go a long way. We should also keep in mind that these negotiations aren't always zero-sum. Often the parties have distinct but not opposing interests for many typical terms. The key to success is identifying what's important to each party early. This can shorten negotiation time and improve satisfaction of the parties by limiting or avoiding conflict on terms where neither party has a strong interest or where only one party has a strong interest that is not a significant detriment to the other party.

I just participated in a University study on this topic. Professional negotiators were put into 3 groups, 2v2 with OUT ChatGPT, 2v2 BOTH using ChatGPT, and 2 without v 2 with ChatGPT. Both "equal" groups ended with a fairly equal deal, but the group with OUT ChatGPT did significantly better than BOTH their counterparts, who used ChatGPT, but ALSO the other 2 groups negotiators who were on the same "side" of the negotiation - by 8-10%. Everyone agreed ChatGPT is useful for planning and recapping, and potentially on a timeout, but in real time it's a very difficult and distracting disadvantage.

AI is also prone to bias (whether intended or unintended). It can only act on what it learns from the sample data provided to it. So, for example, an AI tool that has been fed primarily manufacturing agreements or product sales may not be a good fit for a professional services agreement. Likewise, if your AI tool was trained primarily on contracts executed by parties with outsized leverage (e.g. Microsoft, JP Morgan, Walmart), then the tool may learn 'standard' position are what benefits parties with huge buying/selling power which wouldn't necessarily fit your needs in most transactions.

AI tools are very good at drawing comparisons and expressing a co.plex concept in a simple way. It will be useful if correctly utilized to bridge the knowledge gap that you have of your counterparty. Used for Non-realtime negotiations, like when marking-up a contract, you can be more persuasive if arguments are expressed in the counterparty's language (e.g. "explain why X should be so to an engineer").

At the heart of AI tools are algorithms that can review contracts and identify certain items more quickly and accurately than manual review. This can save time and reduce errors, allowing you to focus on the more important aspects of the contract. AI tools can also assist with compliance, by detecting potential risks and alerting you to possible inconsistencies.

In addition to the above AI tools will be limited because of the inability to anticipate legal challenges and removes the contracting parties autonomy. This is why it’s imperative to include the humanistic elements during the verification and validation processes.

However, AI tools also have their drawbacks. They may not be able to identify all of the nuances in a contract and may overlook subtle differences. Additionally, AI tools are not able to provide legal advice, so you will still need to review the contract for accuracy.

If corporates are truly interested in "time savings" to the extent they're considering delegating the negotiation of certain contracts to an LLM, they're better-off considering alternative methods such as: (1) seeing if a biz case could be made to exempt certain contracts from detailed review altogether, e.g. contracts with a contract value <$x, with LoL/indemnities not exceeding 300% of contract value, no reg. requirements, etc. If the contract owner confirms that commercial terms are accurate and all "whitelist" criteria are met, it might not be necessary to spend time on complicated reviews/approvals across the org. (2) supporting the use of standardised contracts across the industry(ies), eg with the use of things like oneNDA

When implementing AI tools it is important to consider the intent of the creator of the AI tool. What values did they include in the tools during the development of the AI tool. Did they design an AI tool that looks to protect the individual autonomy or the company bottom line. Does the AI tool take into account the varying legal obligations across multiple territories, and thereby giving a fair and balanced view or was it built to take solely into account a legal standpoint from a single territory, which may be less favourable to a contracting party.

Was this written using ChatGPT - it is more of a summary of the issues than an analysis. I used ChatGPT recently to provide a contract review and it missed the strategic issues - part of this omission was likely due to lack of information but I would only use AI to assist right now as it is excellent at giving you a starting point. All AI content produced by ChatGPT needs to be checked as it is a language model and its information sources are therefore based on probability and fallible