prediCtive anaLytiCs in LegaL praCtiCe: innovation with intentionaLity
Technology Section
Continued from page 51
she explain the basis for a recommendation to a client? Professional responsibility and ethics rules require and place great emphasis on informed consent and clear client communication— this is challenging to achieve when the basis of a recommendation is concealed within a proprietary algorithm.
3) Overreliance on Probabilities Every experienced legal practitioner knows there simply is no slam-dunk guarantee in the law. While a model might indicate a 75 % chance of success, there simply is no way to input enough data to reduce the outcome to a single number. Predictive models might consider historical cases, judicial behaviors, fact patterns, but can they account for the intangibles like witness credibility? Lawyers( and clients) must take care not to treat probabilities as certainties, or risk strategic missteps, poor settlement decisions, or misaligned client expectations.
4) Data Security
As always, with any predictive( or generative) technology, lawyers must consider the security of the program being used before feeding it sensitive client information. Without strong data governance and security, the risks of exposing privileged material or violating confidentiality obligations become unacceptably high. Law firms are already high-value targets for cyber threats; utilizing predictive modeling and other similar technologies only increases the need for rigorous and up-to-date security practices.
Predict Responsibly: 1) evaluate the Data and the Model Lawyers should vet predictive tools with the same scrutiny applied to expert witnesses. Essential questions include: What data sources are used? How is bias mitigated? How often is the model updated? What assumptions drive the predictions?
2) Center human Judgment The IBM Training Manual from 1979 contained the following quote:“ A computer can never be held accountable, therefore a computer must never make a management decision.” Predictive analytics should support— not replace— legal reasoning. Firms should establish guidelines clarifying how predictions are used, how they’ re communicated to clients, and how they fit into broader strategic analysis.
3) expand Governance and Training All lawyers should already have data security and management policies. But with new technologies comes new responsibilities. Lawyers must ensure their firms regularly evaluate and update policies around data use, retention, disclosure, and model performance review and that all lawyers and staff are properly trained.
4) educate Clients and Set Reasonable expectations While clients increasingly expect their counsel and support staff to utilize time-saving technologies like predictive modeling, they often lack a full understanding of the risks and limitations involved. It is therefore incumbent on the lawyer to understand these technologies well enough to communicate when they are reliable, when they are not, and the reasons behind those distinctions. n
Author: Caroline Spradlin – Phelps Dunbar LLP
Attorneys Needed for HCBA Lawyer Referral Service
The HCBA would like to extend an invitation to all attorneys to join the HCBA’ s Lawyer Referral Service. Bilingual attorneys are especially in need as the local Spanish-speaking population is underserved in the following practice areas: probate, consumer protection, immigration, landlord / tenant and business.
To learn more and to join the HCBA’ s Lawyer Referral Service, visit www. hillsbar. com / page / JoinLRIS..
Contact Lupe Vazquez-Mitcham at 813 221-7783 or lupe @ hillsbar. com for further information.
M A r- A p r 2 0 2 6 | H C B A L A W Y E r
5 3