Praful Saklani is the co-founder and CEO of Pramata, an end-to-end contract management solution. Views are the author’s own.
The promise of generative AI is bewildering.
In a matter of months, from the launch of OpenAI’s ChatGPT-3 last November to the GPT-4 release in March, the technology is rapidly shifting from being a novelty for most to becoming an essential component of our personal and professional lives.
As someone involved in the tech world since the mid-1990s, and within the legal tech space since 2005, the only similar epochal leaps I can recall are the launch of Netscape’s browser that opened the Internet up to everyone in 1994 and the release of iPhone 3 that included the addition of the Apple’s App Store, GPS-tracking, a significantly improved camera and 3G-wireless capabilities.
In both cases, these technologies passed the ‘8 to 80 test’ – meaning it was intuitive for everyone regardless of their age. It was just a matter of time before the technologies changed the lives of children, their parents, and their grandparents.
As of today, generative AI is officially on the same path.
What is even more astounding is just how quickly advancements are happening. If I had written this column earlier this year, prior to the launch of GPT-4, I would have predicted generative AI may have an impact on our industry in the next three to five years.
As the technology stands today, I have no doubt it will quickly become an indispensable tool for any attorney in the contract management space.
Generative AI’s immediate impact on contract management
Because generative AI is easy to use and allows for the creation of very sophisticated experiences with almost no coding required, it will soon change SaaS applications and legal tech in ways we cannot begin to imagine.
What does this mean for the legal community and contract management teams?
For starters, even at this early stage of generative AI, the technology could potentially serve as an assistant to an experienced attorney.
It has already reached a startling level of accuracy, capable of performing a wide variety of research work that can then be reviewed by a trained legal professional.
According to OpenAI, its GPT-4 technology not only passed the bar exam, but scored in the top 10% of test takers. (By comparison, OpenAI’s ChatGPT-3 scored in the bottom 10%).
As it currently exists, I wouldn’t trust GPT-4 to understand every legal nuance, but I think it can minimize a major amount of research time and offer very helpful suggestions in the hands of a well-trained attorney.
Right now, the vast majority of interaction with generative AI is through OpenAI’s ChatGPT application.
While this is an easy to use platform that is already being accessed by millions of people every day, OpenAI’s plugin vision allows other software vendors to build their own plug-ins and APIs which can integrate the technology into any web-enabled application and vice versa.
This means ChatGPT will eventually be able to use any web-enabled application, and any web enabled application should be able to harness GPT!
From a legal tech and contract management perspective, the most anticipated GPT plugin is Microsoft’s integration of its GPT Copilot into Microsoft Word and Office 365.
This advancement has the potential to massively improve contract template creation, document creation, and legal review aspects of contract management. In my view, this move by Microsoft will likely replace much of the document editing and redlining technology inside many CLM platforms.
Imagine being able to tell your AI Copilot inside Microsoft Word to quickly complete small modifications to contracts in record time or to quickly compare contract terms for multiple contracts.
For example, simple prompts like “Create a renewal order form for Company A with an expiration date of X, using the same format as this previous order” or “Recommend termination language for contract A similar to the language used in this other contract” could end up saving uncountable hours for contract management teams in large and small companies alike.
Generative AI’s confidentiality issue: Are attorney-client privileges at risk?
Before generative AI technology can be used at scale in “enterprise-grade” applications, one major challenge still needs to be addressed: confidentiality issues.
Current generative AI platforms pose serious privacy and security risks when it comes to confidentiality and the threat of exposing sensitive data.
A recent report by the data security platform Cyberhaven showed it had blocked requests to input data into ChatGPT by 4.2% of the 1.6 million workers at companies that use its security software. The reason: the employees were at risk of leaking confidential information, sensitive business data, and source code to the large language model (LLM) that underpins ChatGPT.
The very same circumstance could happen within legal circles. If an attorney or anyone in the contract management space enters contract terms into a generative AI platform, they would be putting attorney-client privileges at risk and potentially breaking confidentiality terms.
My own team’s research into the technology’s security has revealed there are definitive security and confidentiality challenges. So much so that any user should tread very carefully with commercial-grade versions of GPT-4 that claim enterprise-level security capabilities.
That doesn’t mean privacy and security issues will be a long-term problem for generative AI.
OpenAI, Amazon, Microsoft and other organizations have announced they are working on enterprise-level, private cloud LLM options. As current applications continue to advance, so will the technology’s security and safety levers.
The Future of AI
With the launch of GPT-4, we have entered a new era of generative AI that will impact every part of our lives in the next two-to-three years.
And while the technology’s pace of advancement is truly dizzying, GPT still follows the same guiding principle of other technologies – that what you get out of it wholly depends on what you put into it.
From a contracting perspective, contract management teams still need to have copies of signed contracts in one place. They still need to maintain high levels of contract data accuracy, and they need to know which contracts are active and to which accounts they belong.
They will continue to need coherent contract management policies and playbooks that reflect their business’ risk tolerance and objectives.
These fundamental contract management components will continue to be essential even in an AI-enabled world, and GPT does not change that.
If your team is unable to give GPT the correct contracting inputs that synthesize correct information and generate accurate content, generative AI won’t be useful or effective for even the most mundane tasks.
How the technology will evolve within the legal industry remains to be seen.
For now, the one thing I know for sure is that wherever generative AI is headed, it is going to get there much faster than any of us realize. My hope is that it continues to be just as awe-inspiring as what we’ve experienced so far.