Colorado AI Ethics Rules for Lawyers: What the Bar Says (2026)

The Colorado Bar Association issued Formal Ethics Opinion 145 in March 2024, becoming one of the first state bars to publish a formal opinion specifically addressing AI use by attorneys. The opinion provides clear, actionable guidance on billing, confidentiality, and disclosure — three areas where attorneys face the most immediate risk.

Colorado's early action reflects the state's progressive approach to legal technology regulation. Opinion 145 has become a reference point for other state bars developing their own AI frameworks, and its billing guidance is among the most specific in the country.

What the Bar Says

Formal Ethics Opinion 145 (March 2024) analyzes AI through Rules 1.1 (competence), 1.5 (fees), 1.6 (confidentiality), and 5.3 (supervision). On competence, the opinion holds that attorneys must understand AI tools well enough to recognize errors and limitations before using them in practice. On supervision, AI-generated work product must be reviewed with the same diligence as work produced by a paralegal or junior associate. The attorney is ultimately responsible for every word filed with a court, regardless of whether AI generated the first draft.

Billing Implications

Colorado's opinion is direct: attorneys must not charge clients for AI tool costs at attorney hourly rates. AI subscription fees may be passed to clients as costs or expenses, but they cannot be embedded in hourly billing at attorney rates. If AI reduces a 5-hour research task to 30 minutes, the client should see the benefit. The opinion requires transparency — clients must understand what they are paying for and how AI affected the work. This is one of the clearest billing positions any state bar has taken, and it sets a standard that other states are adopting.

Confidentiality Rules

Opinion 145 requires attorneys to vet AI tools for data protection before using them with client information. This means reviewing terms of service, understanding data retention policies, and confirming that inputs are not used for model training. The opinion does not list specific approved or prohibited tools but places the burden on the attorney to conduct reasonable due diligence. Attorneys who use AI tools without understanding their data practices violate Rule 1.6 regardless of whether an actual data breach occurs. The duty is prospective, not reactive.

What's Still Unclear

The opinion recommends disclosure when AI was substantially used but does not define the threshold for "substantial" use. Running a single AI spell-check differs from having AI draft an entire brief — the line between these is left to attorney judgment. The opinion also does not address AI use in specific practice areas, CLE requirements for AI competence, or the interaction between AI ethics and Colorado's consumer protection laws. Colorado has not yet addressed whether AI governance policies are required at the firm level, though the opinion strongly implies they should exist.

The Bottom Line: Colorado's Formal Ethics Opinion 145 set an early national standard: vet your AI tools, protect client data, bill transparently, and disclose substantial AI use.