New research suggests that fine-tuning AI models can indeed be worthwhile, even when using retrieval-augmented (RAG) setups—but only in specific cases. While fine-tuning large models hasn’t consistently proven valuable for legal applications, recent findings show that focusing on smaller models can yield substantial improvements, especially in tasks like information retrieval and accuracy in legal work.
Nicola Shaver
LTH Expert Legaltech Hub
Nicola Shaver is the CEO and co-founder of Legaltech Hub. She has 20 years of experience in the legal industry, including ten years of practice experience with top tier firms and Fortune 500 companies and close to a decade of global experience as a senior innovation leader with international firms.
Her highly innovative approach to legal business transformation, leveraging best practices from outside of industry and building capabilities such as client-facing teams and technology products, led to recognition by ILTA as Innovative Leader of the Year in 2020, the same year that her firm was named Innovator of the year. In 2021, she became a Fastcase 50 honoree and a Fellow of the College of Law Practice Management.
In addition to her work with Legaltech Hub, Nicola is an adjunct professor at Cardozo Law School, where she teaches the school's inaugural course on legal technology. She is a frequent advisor to law firms, corporate legal departments and legaltech companies and has been invited to speak at conferences in Australia, Canada, the United States, Latin America, and Europe on topics such as digital transformation and technology adoption. She is a regular contributor to publications including Law360, ALM Legaltech News, Modern Lawyer, and Legal Business World, and a passionate advocate for positive change in the legal industry.
Loading...