"Interview with the AI" - a discussion of legal use cases with the source itself

The legal community is racing to take advantage of the best use cases to produce efficiencies using generative artificial intelligence (GenAI). ⁴  So, during my interview, I explored the emerging legal use cases with the source itself, the free public ChatGPT application.

For this article, I have attributed all content to the publicly available ChatGPT application by taking direct quotes from the tool’s responses with zero embellishment -- edited only for brevity. Comments contributed by me appear under my name exclusively. Any of you who have tried ChatGPT will quickly recognize that the “questions” presented in this article are crafted in a conversational “interview” style. The actual queries used with ChatGPT are, by necessity, often stilted and iterative and make for far less entertaining reading.

So, without further delay and with a reverent nod to horror genre author Anne Rice’s “Interview with a Vampire”, I present, “Interview with the AI: a discussion of legal use cases with the source itself”.

Diane:  We are sitting down today with ChatGPT, to explore the use cases that GenAI – like OpenAI⁵-- will be solving for legal. Let’s dive right in. ChatGPT, where do you see the first big opportunity for GenAI to contribute to legal?

ChatGPT: “The process of drafting contracts can be laborious and time-consuming, often requiring significant attention to detail. OpenAI's⁶ potential in contract drafting can be understood through its capabilities in template generation and customization.

By leveraging its Natural Language Processing (NLP)⁷ capabilities, OpenAI can generate and customize contract templates based on a few simple keywords. Remember that one time you overlooked a critical clause in a contract? Its keen artificial eye can spot inconsistencies and potential issues that might otherwise slip through the cracks, ensuring that contracts are watertight, and parties are protected.”

Diane:  That sounds amazing, but are there any concerns?

ChatGPT: “It is worth noting that these capabilities have not been widely deployed yet, so a degree of caution is warranted.”

Diane:  Yet you see contract drafting as having significant potential, where do you see it going?

ChatGPT: “OpenAI is already flexing its muscles in drafting simple contracts and agreements. Whether it's a straightforward non-disclosure agreement or a vanilla sales contract, Open AI is making its mark.

However, as powerful as OpenAI's capabilities might be, it is essential to maintain legal compliance and human review.”

Continuing role for humans

Diane:  Wow, I’m surprised ChatGPT is advocating for a continuing role for humans in this process.  And, as a human and a lawyer, it’s reassuring to hear. The use of GenAI for legal is in early days. We’ve all experienced the heavy lift of training existing (arguably now legacy) AI models requiring large or refined sample contract data sets for iterative learning that were often disappointing.  Those days are soon over as the burden shifts from worrying about precision and recall to setting “the temperature” of your AI inquiry. It means controlling your prompt structure and data feeds to reduce bias and fiction generation and to fine tune responses.

Tell us in summary ChatGPT, why do you believe humans will still be important to contracting?

ChatGPT: “Trusting the technology is one thing, but ensuring the final output is legally sound is of utmost importance.”

Diane:  That’s an excellent sound bite ChatGPT.  What other use cases do you see emerging for legal contracting?

ChatGPT: “Contract review can be a painstaking process, often plagued by human error and oversight. As the technology evolves, it is likely to become more adept at reviewing specialized contracts and providing risk assessments for even the most niche industries.”

Diane:  I agree with you in part. Providing contextual assessments of contract changes is an area where GenAI will impress users. But true risk assessment requires a qualitative, subjective understanding of the unique needs of the relevant business. Humans will be needed to calibrate, update, and contextualize risk estimates. Formulas and algorithms will only go so far.

What industries have the greatest potential for early value in your estimation ChatGPT?

Chat GPT: “The technology, real estate, and employment sectors have relatively standardized contracts, making them ideal candidates for early success with OpenAI applications.

As OpenAI continues to evolve and mature, industries with more complex and specialized contracts, such as finance, healthcare, and intellectual property, could see significant long-term value.” 

Diane:  Here we differ ChatGPT. I see expansive early potential in finance and healthcare. These industries have consistent communities, and their fair share of repetitive processes and standardized templates for numerous activities. Those areas are ripe for intelligent automation with a key role for human quality control and process governance.

Let’s move on. You’ve called out drafting and contract review as viable use cases but there must be more potential?

ChatGPT: “As OpenAI continues to evolve, it could potentially provide predictive analytics and contract intelligence, offering invaluable insights to businesses.”

Diane:  This all seems too good to be true, what should we be thinking about in taking these capabilities mainstream?

ChatGPT: “Some key concerns include legal and ethical considerations, ensuring data privacy and security, maintaining a balance between automation and human expertise, potential bias in AI-generated contracts, and the role of regulation and standardization.”

Diane:  Let’s take those concerns individually.  Arguably law makers and regulators are already late to the game. 

• UNESCO has made a general plea for good behavior by adopting a recommendation on the ethics of AI.
• Proposed rules by the Cyberspace Administration of China require AI content to adhere to that country’s socialist values.
• The EU is making progress with the AI Act, a regulatory scheme that focuses on controlling what it defines as unacceptable or high-risk applications.
• In the US we can keep an eye on The National Institute of Standards and Technology as a source of guidance, but we cannot afford to rely on court interpretations of existing frameworks to be agile enough to address what will be issues of first impression.
• Law firms and legal leaders must do proactive work to anticipate areas of concern and prepare responses in support of their business clients who are adopting GenAI now to be competitive. That work will then inform the slower moving regulatory framework that will follow.

Ok, ChatGPT, let’s talk about the concern you raised around bias. Psychologists have shown humans hold innate and learned biases. Since human information is what GenAI ingests, do we pass those biases along to the AI?

ChatGPT: “AI systems, including OpenAI, are trained on large datasets and can potentially inherit biases present in the data. Ensuring that AI-generated contracts are fair, unbiased, and representative of the parties' intentions requires continuous monitoring and improvements to the underlying algorithms.”

Diane: Please elaborate on the other big concern you mentioned, how to assure data privacy and security?

ChatGPT: “Data privacy and security are of paramount importance in the digital age. As OpenAI processes sensitive contractual information, robust measures must be in place to protect data from unauthorized access, misuse, and breaches. Additionally, businesses and legal professionals must adhere to relevant data protection regulations, such as the General Data Protection Regulation (GDPR).”

Diane:  Here ChatGPT’s response essentially restates the question, and it illustrates a core issue we face with this technology. GenAI is only as good as the questions we ask it. It will be incumbent on organizations launching with this technology to engage with vendors -- people who can operate seamlessly and effectively with the query nuances required to get consistently meaningful and accurate content.

Also needed are vendors -- people who know what robust security measures must look like and how to deploy them to protect privacy and other interests. Concurrently, it will be critical for legal to learn what exposure this technology creates and articulate for our businesses what satisfies existing and anticipated regulatory requirements.

Diane:  One final use case we didn’t discuss is contract negotiation, an area many in the legal profession will say is an artform outside the realm of automation.  How would you respond to that thinking?

ChatGPT: “Negotiating contracts can be a complex and time-consuming task, often requiring a deep understanding of the nuances involved. OpenAI's potential in contract negotiation lies in its ability to analyze negotiation strategies and tactics, as well as its capacity to automate response generation and potentially assist in complex negotiations and multi-party agreements. However, it is important to approach this area with caution, considering the ethical implications and the need to preserve human intuition in negotiation processes.”

Diane: Facilitating the interplay of parties, providing transparency on positions, exposing historical choices, clarifying clause meaning, and simplifying wording will all be extraordinarily valuable additions from GenAI to the negotiations process. Fine-tuned applications that optimize GenAI use for specific industries and reduce bias and consistency concerns will further make negotiations a better experience with GenAI in the mix. 

Yet, remember Contracts represent relationships and customer/supplier/partner satisfaction is a component while political and strategic overtones will color desired outcomes. Certainly, these considerations factor less in routine transactions satisfied with automated low-risk back-up positions. I believe there remains a role for humans to ultimately form the bonds and color the relationships a contract represents.

And to that end, I’m heartened that you, ChatGPT, believe there is a place for human intuition in the negotiation process!  This is fascinating. What role do you see humans playing in all of this?

ChatGPT: “It is essential to remember the importance of expert legal oversight. OpenAI, as remarkable as it may be, is still a machine and can inherit biases from the data it's trained on. Ensuring a balance between automation and human expertise is crucial. As enchanting as OpenAI's capabilities may be, it's crucial to maintain legal compliance and human review.”

Diane:  Is there another more compelling way to say that?

ChatGPT: “After all, we don't want our AI overlord to inadvertently put us in a sticky legal situation. As the old saying goes, "trust, but verify."

Diane:  In the above question I changed my approach slightly and asked ChatGPT to take a different tone.  In response, it assigned itself the identity of “AI overlord.” That is interesting because we humans tend to personify technology.

ChatGPT with all its bravado is not a person and we cannot forget it. GenAI responds to where we lead it today (though, to be fair, so do some people). Does ChatGPT truly “believe” there is a place for humans in the contracting process? No, it is echoing the sources it has trained on and is responding to the directional bias of my questions. But those of us humans looking at this amazing disruptive technology believe there is a place for us.

The focus and role for legal will be to make sure the touchpoints GenAI creates throughout the businesses we support, are understood and reliable. And we cannot shy away from embracing challenges in areas like negotiation just because we think it cannot be automated. Closing your eyes is not going to make the AI overlord go away. Accept that it is here, work to mitigate the dangers, and use its powers to your advantage. ChatGPT, any final words of wisdom?

ChatGPT: “The potential of OpenAI to revolutionize contract drafting, review, negotiation, and analytics is immense. By addressing legal, ethical, data privacy, and security concerns, and maintaining a balance between automation and human expertise, OpenAI could become a valuable tool in the world of contracts. The future of contract management is evolving. So, let's embrace the innovation…”

Diane:  Thanks, ChatGPT, that was informative, just a bit chilling, and fascinating to get your expert opinion.

Final thoughts for our audience: contract generation, drafting assistance, contracts review, analytics, and yes, even negotiation are all legal use cases for GenAI that are absolutely happening. ChatGPT itself just confirmed it and with a fair degree of humanlike persuasion.

Our job as actual humans is to understand, query, and quality check the technology effectively. There are developers out there who are incorporating this technology in responsible ways. There are partner organizations who are skilling-up to run these environments and deliver services that effectively use these new capabilities for optimal efficiency. The time to learn is now, the time to act is coming fast.

Will AI ever not need us? That is an interview for another day.

ABOUT THE AUTHOR

Based in Cleveland, OH, Diane is a 30 year in-house legal practitioner and former WCC Board Chair now with Integreon  She has spent the last six years in the alternative legal service provider space supporting clients and services teams in applying technology to solve legal environment challenges. A published thought leader, Diane has worked with a broad range of legal technology spanning contracts repositories, AI, workflow management, and reporting analytics. Diane published an article titled Enhancing relationships put people first – a story of success! in the CEJ on 29 April 2022.

ABOUT INTEGREON

Integreon is a leading provider of outsourced legal and business services empowering corporations and law firms to meet today’s complex challenges and achieve more.

END NOTES

1. 21 ChatGPT Use Cases: Save Time and Increase Productivity, Enterprise DNA, AI, ChatGPT blog series
2. What is CHATGPT and why does it matter?  Here’s what you need to know, Sabrina Ortiz, Associate Editor, ZDNET Innovation.  Refer also to ZDNET article titled, How to use ChatGBT: Everything you need to know, Maria Diaz, June 30, 2023)
3. Chatbot definition, Wikipedia
4. Generative AI articles.  Reference: Generative AI in law firms: many such technologies are still a great unknown, Zach Warren, May 23, 2023 Reuters. Refer also What is generative AI? Everything you need to know, George Lawton, Tech Target, Enterprise AI, Tech Accelerator
5. Generative artificial intelligence (GenAI) and OpenAI, private research laboratory founded by Elon Musk
6. OpenAI article (reference only – not for advertorial purposes)
7. Natural Language Processing (NLP) Wikipedia definition