California and the IRS Turn AI Controls into Practice Rules

Legal AI is moving from experimentation to governed practice, with verification, confidentiality, and accountability becoming mandatory parts of the workflow.

Updated

Part of a broader trend

AI’s Confidence Illusion: How Overtrust, Hallucinations, and Lax Oversight Are Tripping Up the World’s Biggest Firms

Fluent AI is winning trust faster than it earns it—and that gap is becoming a business risk.

Part of a broader trend

CFOs Race to Master AI as Automation Transforms Finance—But Execution Jitters Persist

Finance teams are moving from AI pilots to core automation, but the hard part is making it reliable.

Part of a broader trend

AI Ambitions Outpace Reality: Enterprises Stumble on Data, Security, and Trust as Agentic Rollouts Stall

AI spending is surging, but enterprises are hitting a wall on data, controls, and trust.

Part of a broader trend

Courts, Lawmakers, and Regulators Redraw the Liability Map for AI—With Google and Legal Pros on the Hook

AI liability is shifting from the machine to the companies and professionals behind its outputs.

What is this trend?

California and IRS rules are turning AI use in legal and tax work into a controlled, reviewable workflow where lawyers must verify outputs, protect client data, and stay accountable for every AI-assisted filing or advice.

  • Confidential client data can’t be fed into public AI tools without controls.
  • Lawyers must verify AI outputs and remove hallucinated citations before use.
  • AI-assisted work remains the lawyer’s responsibility under ethics and tax rules.
  • Courts are sanctioning failures to review AI drafts as conduct violations.
  • Demand is rising for auditable AI systems with traceability and sign-off.

What’s the latest?

California’s SB 574 would make AI use a controlled workflow requirement: lawyers and arbitrators could not place confidential, personally identifying, or other nonpublic client data into public genera

AI Controls Become Practice Rules, Firmwide Adoption Accelerates, and Taiwan Advances AI Governance

How it developed

  1. AI output verification becomes core legal skill, legal workflows become governed AI operating systems

  2. AI supervision becomes billable legal work, and AI-embedded workflows push fixed fees

Go deeper

Curated long-form picks on this trend — podcasts, videos, and analysis, by seniority.

Related reporting

Deep-dive stories that report on this trend.

Related trends

Stay ahead in Legal

Get the weekly Legal brief in your inbox — the developments, what they mean by seniority, and what to do next.