The New Standard for AI Review: Defensible by Design
The conversation around AI in eDiscovery continues to evolve, and one recent court decision makes an important point: the problem isn't AI-assisted review—it's whether your workflow is defensible.
In Schulte v. LinkedIn, Magistrate Judge Laurel Beeler's June 30, 2026 discovery order reinforces what many in the industry have been saying for years. Courts aren't asking parties to avoid AI. They're asking them to use it responsibly.
The court rejected attempts to require LinkedIn to run AI review across every custodial file, denied demands for extensive audit-level details about Relativity aiR, and refused requests for broader collections that lacked evidence they would uncover meaningful, non-duplicative information.
Why?
Why?
Because discovery isn't about collecting everything possible. It's about conducting a reasonable, proportional, and defensible process.