Lawyers Need AI With Confidentiality Built In, Not Bolted On: AI That Goes Deep Into the Tech Stack
Lawyers should prioritize AI solutions with built-in confidentiality and governance, rather than add-on tools that create fragmented and risky workflows.Summary
The article argues that lawyers need to demand AI solutions that are integrated into the core of their existing technology infrastructure, prioritizing confidentiality and control. Current AI tools often function as add-ons, creating “shadow workstreams” with inconsistent governance, audit trails, and retention rules, which can lead to legal errors and fines. This fragmentation is particularly problematic given the legal profession’s reliance on maintaining control of information for privilege, work product, and compliance.
The author contrasts this fragmented approach with a “cognitive core” – AI that inherits existing security permissions and governance policies. This embedded approach strengthens compliance, reduces risk, and allows firms to rely on established controls rather than creating new, complex policies. It also simplifies audit trails and reduces version control issues.
Ultimately, the article suggests that the future of legal AI depends on whether firms treat AI as fundamental infrastructure or as a superficial addition. Building AI into the core of legal work, rather than bolting it on, will enable firms to leverage the benefits of AI while maintaining the necessary levels of trust, control, and defensibility.
(Source：National Law Review)
