Last month, we presented to a legal defense group on the evolving role of artificial intelligence (AI) in legal practice. There has been an explosion of technological advancements in recent years, but also an increase in dialogues about how AI will impact work in the legal community. Before we can dive into the specifics of how AI can change any industry, it's first helpful to take a step back and provide a primer on AI.

Below are some takeaways from our research and various discussions with those in the legal community:

While many correlate the birth of AI to late 2022, when ChatGPT was launched, the concept was actually first discussed over 60 years ago. The origins trace back to a 1956 Dartmouth Conference, when researchers described their vision of machines capable of mimicking human intelligence. In the years since, several AI tools have entered everyday workflows, including search engines like Google and digital assistants like Siri. 

However, there are many forms of AI. Using a very general classification, one could break AI down into two categories: traditional AI and generative AI. 

• Traditional AI: analyzes data, identifies patterns, and makes decisions or predictions based on pre-defined rules. It is like a chef following a recipe with precision, following all the rules.
• Generative AI: creates new content (text, images, music) based on training data. It is like a chef experimenting in the kitchen without a recipe — pulling from inspiration to add in flavors or ingredients to enhance other ingredients to make a flavorful dish based on knowledge gained from past experiences.

Based on these definitions, many in the legal community have been using traditional AI for years. It is the introduction of generative AI that has created new challenges and opportunities. Understanding how these tools work and their potential applications is essential for legal professionals looking to integrate AI into their practice effectively.

On the positive side, generative AI can be helpful to:

• Summarize articles and legal decisions into digestible content
• Distill information about an industry in a short paragraph
• Analyze a data set and provide quick bullets about the information presented
• Collaborate with an end user about a subject matter, enabling users to interact with the program to learn more intuitively

While AI can streamline legal research and automate processes, it also comes with risks. One is the issue of AI hallucinations, where models generate misleading or inaccurate information. In a 2024 study published by Stanford, "the Lexis+ AI and Ask Practical Law AI systems produced incorrect information more than 17% of the time, while Westlaw’s AI-Assisted Research hallucinated more than 34% of the time".

This means that AI models may provide responses to inquiries from users that are not true. Therefore, when using generative AI, it is always important to review responses and verify their content. If AI is providing citations, make sure that you review the referenced material. If AI is providing a summary that you want to then use in a legal brief, consider reading the actual underlying document to verify the summary was correct. 

Another challenge of generative AI is how to use it in a way that protects data. Many generative AI models are public and grow based on the information provided to them. Any time a user performs a search or feeds information to the system, that data becomes part of the model. Therefore, it's important for users to not use any privileged information when interacting with public AI models. Once internal company data is put into one of those searches, that information is part of the public sphere and the user that provided it could be subject to confidentiality violations.

While AI presents transformative possibilities for legal practice, it requires the proper oversight to manage limitations and hallucinations. We appreciated the opportunity to share insights on this rapidly evolving area and discuss its impact on the legal profession. KCIC remains dedicated to helping professionals navigate the intersection of technology and law, and we look forward to continuing these conversations.