We’ve been paying close attention as generative AI, thanks to ChatGPT, has exploded onto the market as the next frontier of artificial intelligence (AI). Generative AI is a rapidly evolving technology that has shown great promise in many areas, including the legal profession. However, as with any new technology, implementation has pros and cons. In this post, we’ll explore some of the advantages and disadvantages of using generative AI for legal that we’ve seen to date:
Pros of Generative AI in Legal
As with other forms of AI, the main benefit of generative AI is heightened efficiency. Generative AI can analyze and process large volumes of legal data and documents at a much faster rate than humans, offering summaries and suggestions on legal language. This saves legal professionals significant amounts of time and energy, allowing them to focus on more complex tasks.
Utilizing generative AI is predicted to bring major cost-savings to reduce spend on outside counsel. While the use of generative AI may be perceived as an avenue to replace jobs in the legal sector, there is a high probability that it will actually increase in-house hiring as it mitigates the need for costly outside counsel spend. There are still ample areas of legal responsibility that are not suited for AI as they require a great deal of human interpretation.
Agile teams looking to use the hottest technology will certainly outpace those still completing manual, repetitive tasks. In a world of constant AI innovation, employing the latest tool and demonstrating that you are ready to adapt your workflows to technology instead of being left behind can be exciting.
Cons of Generative AI in Legal (at this point in time)
Artificial intelligence algorithms are only as good as the data they are trained on. If the training data is biased, the AI could amplify existing inequalities or prejudices. More specifically, generative AI models can “hallucinate” wherein they produce confident-sounding responses that are actually incorrect. This leaves users vulnerable to perpetuating false information. As such, this technology is currently more useful as a supportive assistant rather than a replacement to any high-risk business processes that require human reasoning and judgment.
Data Privacy Concerns
Rapid advancements in generative AI are outpacing the regulations and research needed to ensure this technology is being used ethically. Tools like OpenAI’s ChatGPT retain user input data to further train their models. In the event of a data breach, which has already happened with OpenAI, any proprietary information users have input could be leaked. Companies rushing to integrate with generative AI APIs may be exposing their customers to significant risk ahead of evidence of the long-term implications of this technology.
Lack of Human Oversight
The power of generative AI may appear to some users as almost magic. But it’s worrisome if users become complacent in quality control when relying on this form of AI to streamline their work. Without visibility into the types of training data and authors of source material, it can be difficult to rely on this technology to adapt to unique legal needs. As generative AI is sourced from large data pools vs. training data fine-tuned to legal language, results may require additional tweaking and human input to get high-quality results.
While implementing generative AI in legal has many potential benefits, it also raises some serious concerns. As advancements have outpaced regulations on this technology, it’s important for legal professionals to carefully weigh the pros and cons when using generative AI or assessing a vendor that employs this technology.
It’s less of a question of whether generative AI will significantly impact legal and more of a question of when and how is the appropriate time to leverage these tools. Advancements in this space are coming in every day. Subscribe to the LinkSquares blog to stay in the know.
Subscribe to the LinkSquares Blog
Stay up to date on best practices for GCs and legal teams, current events, legal tech, and more.