Artificial intelligence (AI) has been a topic of discussion for a long time. The main points that are seemingly always brought to the forefront is how AI would look in all fields of industry, what safeguards would be in place as preventative measures, and what is the limit for the AI itself. In recent years, AI has found itself to be an innovative way for individuals to structure their workflow and alleviate certain tasks that might take a longer amount of time to complete. While AI can be used in place of certain functionalities, it is important to discuss concerns and measures that placate those concerns.
One of the biggest concerns that individuals or firms may have, especially when it comes to AI in the legal field, is the idea of customer data being used to teach any AI how to better serve the user. It is up to legal tech firms to create preventative measures to protect customer data and enhance their AI tools all at the same time. The necessity for legal tech providers to protect customer data can be considered as the most important piece when it comes to building or enhancing data. With that in mind, it is equally as important to build AI tools that can compete with other tools that may be out there, but how do you do that without the use of test data? By having a development team that knows how to enhance the tool by using data sets that are still going to help the tool understand what to look for but without compromising any customer data. Having strong teams that focus on this is vital for any company that is offering AI.
Another issue that may deter firms from adopting AI tools is the idea that, somehow, any work created by AI will not be approved by an actual human. In other words, that there is no way for human auditing to happen on AI work product. This was seen in the Mata v. Avianca case, where a lawyer used ChatGPT to research their brief and failed to audit the results of the research. The LLM hallucinated fake case extracts and citations, and the lawyer filed them with the court. Due to such risks, state bars are starting to clarify ethical requirements when it comes to the usage of AI in litigation. (For example, the State Bar of California issued practical guidelines when it comes to generative AI. You can read more about that on our blog here.) Companies that provide AI tools must understand that, while AI is there to supercharge productivity, it is of extreme importance to have human audits on the work product generated by AI. Without this, there would be a lack of ethical limitations. There are certain ways that firms can implement human audits when it comes to AI usage: having the user read over the work product created by AI before finalizing the product, allowing users to edit or modify the work product, and showcasing which pieces of work have been touched by AI, whether it be strict AI work product or an AI-human collaboration.
With more and more platforms including AI in their workflows, it is important for users and companies alike to research and understand the potential impacts of AI and what safeguards should be considered when integrating AI into their workflows. Outside of this, it is also important to acknowledge how revolutionary AI can be to any field. Speaking of just the legal field, AI tools can help shave off hours of time, allowing for individuals to work at a faster and more efficient pace, leading to a more cohesive and collaborative workflow. So, while it may be beneficial to remain slightly cautious of newer pieces of technology, it will also be beneficial to understand how a legal tech provider enhances their particular AI tool and what safeguards they have in place to protect users.
