AI

Litigation is now digital, but is it more efficient?

Don’t “chat” with AI like a junior lawyer. Long conversations degrade results. When output slips, start a new chat: rephrase, add context, or break the task into smaller steps.

LLMs summarize well but default to average analysis. That makes them useful for baseline work, but not a substitute for the fresh, above-average insight junior lawyers often provide.

Delegating legal work to LLMs can save time but costs learning. Prefer: you read/outline; AI summarizes and fills tedious parts. Ask not only “can AI do it?” but “am I OK not doing it?”

LegalTech favors “direct” AI control, but its benefits are shrinking as frontier LLMs improve at prompting, workflows, and security. The “inverted” model is viable and likely to gain adoption despite vendor resistance.

Frontier AI models improve so fast that studies comparing them to legal-specific AI become outdated almost immediately. Retesting recent benchmarks shows general models already match or exceed earlier reported gaps.

Lawyers’ claims that AI makes them more efficient may be unreliable. Evidence shows people often misjudge AI’s impact on their own productivity. Firms should rely on objective data, not self-reported impressions, when de