Unless you have been living under a rock, you’ve probably figured out that Chat Gpt might be an awesome solution for the complex and time-consuming process of drafting contracts. I use the word “might” because some serious contract drafting experts are warning to stay away from Chat Gpt.

For example, Ken Adams provides the following advice: “If you’re looking for optimal drafting, the answer is easy: stay away from ChatGPT in particular and artificial intelligence in general.” (If you don’t know Ken Adams, he is a drafting guru with all sorts of great insights and resources for contract drafting.)

But, it’s hard to ignore ChatGpt when it’s just so damn amazing at creating beautiful contract-sounding language. Even worse, what if your competitors are getting so good at using AI for their contracts that they can leave the office by noon, while you find yourself drafting after dark?

If you find yourself trapped by this uncertainty, this article is intended to help you find a way forward. In particular, you can learn in this article about the following:

A demonstration of how Chat Gpt utilizes client instructions for drafting
Some warnings about Chat Gpt’s weaknesses
Some tips about how to use Chat Gpt in your drafting

A Drafting Exercise

When I teach contract drafting to lawyers, I start with the following drafting exercise.

Exercise Instructions: You can find below some client instructions to a lawyer about the modification of a Sales Contract. Based on the instructions, quickly draft a provision (or provisions) for the Sales Contract.

Client Instructions

We need to add some language to our standard contract for selling cars to customers. Several of our customers have been taking advantage of our generous policy allowing them to return cars within the first 30 days after purchase if they are not satisfied with their cars. (In several cases, the customers returned cars that the customers had clearly smashed up.) We want to explicitly state the above policy in our Sales Contract with a qualification that customers will not be allowed to return cars in the event of damage caused by the customer and that the customers need to provide a reasonable reason for returning the car (but no reason is required if the customer merely would like to exchange the car for a more expensive model – and will pay for the price difference).

GPT: Drafting on Steroids

When I gave the above instructions to Chat Gpt, it spit out the following answer:

Section 6: Return and Exchange Policy

6.1 Customer Satisfaction Guarantee: We are committed to ensuring your satisfaction with your purchase. If, for any reason, you are not satisfied with your car, you may return it within 30 days from the date of purchase.

6.2 Eligibility for Return:

(a) To be eligible for a return under this policy, the car must be in the same condition as when it was originally delivered to the customer, subject to reasonable wear and tear.

(b) Damage Caused by Customer: Customers are not eligible for a return if the car has sustained damage or alterations as a result of actions taken by the customer, including but not limited to accidents, collisions, modifications, or negligence.

6.3 Reason for Return:

(a) Customers are required to provide a reasonable reason for returning the car. This reason must be in line with legitimate concerns related to the car's functionality, safety, or performance.

(b) Exception: If a customer wishes to exchange the purchased car for a more expensive model available within our inventory, they are not required to provide a reason. In such cases, the customer will be responsible for paying the price difference between the originally purchased car and the chosen replacement model.

6.4 Return Process: To initiate a return or exchange under this policy, customers must contact our customer service department to receive return authorization. All returns must be accompanied by the original sales receipt and all relevant documentation.

6.5 Inspection and Assessment: Upon receiving the returned car, we will conduct a thorough inspection to assess its condition. If the car meets the eligibility criteria outlined in Section 6.2 and the reason for return is valid as per Section 6.3, we will proceed with the return or exchange as requested.

6.6 Refund or Exchange: If a return is approved, customers will have the option to receive a refund for the purchase price or exchange the car for an alternative model, subject to availability and price adjustments as necessary.

Lessons Learned

With this simple exercise, you can already learn a few key lessons about how Chat Gpt can help and hinder your contract drafting.

Good: It looks amazing

I love using Chat Gpt because it’s unbelievable how good its text can feel. For example, each provision above starts with a helpful orientation heading like the following:

Customer Satisfaction Guarantee
Eligibility for Return
Reason for Return

Also, Chat Gpt has selected a writing style that is more aligned with the reading level of the target users (i.e. ordinary consumers). For example, consider the following comparison between one of Chat Gpt’s provisions and a more “sophisticated” stylistic revision.

Chat Gpt

If a customer wishes to exchange the purchased car for a more expensive model available within our inventory, they are not required to provide a reason. In such cases, the customer will be responsible for paying the price difference between the originally purchased car and the chosen replacement model.

More Sophisticated Style

In the event of the intention to undertake an exchange of the purchased vehicle with a model of greater expense (the “Upwards Exchange”), the provision of a reason is not a precondition to the Upwards Exchange. In the event of an Upwards Exchange, the Parties hereby agree that the Customer shall have the obligation to compensate the Seller for the price differential.

Although the second version above might more closely match the style of certain international law firms, it’s unlikely to put a smile on the face of the client or its customers. Despite Chat Gpt being a virtual robot, it’s crafting text that is more closely targeted towards the stylistic expectations of the intended users.

Bad: It’s misleading

Although I love the look and feel of Chat Gpt generated text, I always find stuff that might get me in trouble if I didn’t read carefully. For example, in the above text, Chat Gpt has created a Customer Satisfaction Guarantee. I find this choice of wording disturbing for two reasons.

First, if I’m a customer, as soon as I hear the words Customer Satisfaction Guarantee, I’m probably going to stop reading and assume that “if I’m not satisfied, there is no way that they will reject my return”. In this case, I might fail to read the subsequent provisions laying out key conditions (e.g. providing a reasonable reason, zero car damage). Such failure could lead to unnecessary disputes and a very frustrated client.

Second, as you know, each jurisdiction has it’s own rules about how consumer guaranties work. (If you don’t know, click here for links about the differing practices just within the EU). Since Chat Gpt is not yet qualified in your jurisdiction, can you really trust it to properly set up a consumer guaranty? Is it going to know all of the codified preconditions to exercising the right? Is it going to be able to tell the client that some of the conditions in its instructions might be in conflict with the relevant legislation? These are important questions that currently only you can answer properly.

Good: Painkiller for Procrastination

We all know the feeling when we need to draft a complex provision but we can’t find some sample language from our firm. In this case, we actually have to draft from scratch, and this is tough because no one ever taught us how to actually draft contracts. (Instead, 99% of us learned by copy/pasting from old contracts and templates.)

If you find yourself in this awkward situation, Chat Gpt is useful because it will definitely give you some text to start with. Although this text is probably far from perfect, you will at least have something to play around with. And, in most cases, this is a wonderful solution for overcoming procrastination (e.g. avoiding work by searching the internet for cute cat memes).

Bad: Says who?

I think we can all agree that a major problem with Chat Gpt is that we have no idea where it got the contract language from. And, source definitely matters for us lawyers. For example, with the above exercise, you probably wouldn’t accept the above Chat Gpt text if I told you that it was based on the standard contract for a 5 car agency in rural Iceland.

And, if you share my feelings about the importance of source, I have a better solution for you. Most of my students don’t know that you can find numerous examples of great contract language from real contracts with really big companies (e.g. Apple, Microsoft). You just need to go to the following link: https://corporate.findlaw.com/contracts.html. Although this alternative approach will require you to find your own language, at least you can be rest assured that your source is most likely an extremely qualified firm.

Good: Stuff is happening

I have a feeling that everything I said above will become moot in a year, but most likely even faster. As global law firms partner with AI companies (e.g. see Dentons partnering with Microsoft) to develop their own internal systems, it’s only a matter of time before we get the ability to tell AI to focus its recommendations using specific parameters like jurisdiction, practice area, industry, client, etc.

At that point, some lawyers are going to lose out on the opportunity to devote countless hours on generating overly-complex documents that the parties rarely read. But, the rest of us can focus our attention on designing and negotiating the key provisions that will make our drafting truly stand out for clients.

Learn More

If you are interested in developing your contract drafting skills, I highly recommend that you check out Ken Adams’s style guide for contract drafting. As an alternative, let us help you with speeding up your drafting process by joining our Academy course on Effective Contract Drafting.

Blog

Quicker Contracts: Pros and Cons of Chat GPT