Why P&T Must Also Govern Algorithms, Not Just Drugs

• A wave of new state laws in California, Colorado, Illinois, New York, and others is creating a fragmented regulatory map for health AI, filling the gap left by federal uncertainty. These regulations focus on data privacy, bias prevention, transparency, and workforce training.

• For pharmacy leaders, this shifts legal responsibility to the health system as the deployer. This means your department may soon be liable for ensuring the algorithms inside your EHR, automated dispensing cabinets, and diversion software comply with conflicting state-by-state standards regarding bias testing and human oversight.

Treat these emerging state laws as an immediate call to inventory every AI-enabled tool in your pharmacy. Assign a governance lead to review vendor contracts for indemnity and bias data, and prepare your P&T committee to oversee software algorithms with the same rigor they apply to the formulary.

• Governance and Task Forces: States like Florida and Tennessee are establishing advisory councils to study AI impact, while legislatures in 30 other states are considering similar oversight bodies.

• Bias and Anti-Discrimination: Colorado has passed the first law explicitly targeting algorithmic discrimination, requiring developers and deployers to perform impact assessments. California and Illinois require attestations that tools used in clinical settings are free from discriminatory bias against vulnerable populations.

• Transparency and Privacy: New laws in California and New York mandate disclosure when patients interact with AI or when AI generates content. Privacy acts in Virginia and Colorado grant patients rights to opt-out of data use or request deletion, complicating how health systems store data for machine learning.

• Insurance and Claims: California now prohibits the use of AI to deny, delay, or modify health care services based on medical need; such decisions must be made by licensed professionals. Washington and Illinois are proposing similar guardrails.

• Workforce Mandates: Illinois legislation explicitly mandates training for health care workers on the ethical use of AI tools. Proposed laws in Massachusetts would require specific informed consent processes for AI use in mental health.

• Governance & Liability: The regulatory shift toward deployer responsibility means Pharmacy and Therapeutics (P&T) committees must actively govern algorithms, not just drugs. You may need to prove that your clinical decision support and antimicrobial stewardship tools have been audited for bias and that your staff is trained to interpret them.

• Operational Workflow: Transparency mandates could introduce new friction. If a state requires patient notification when AI is used, this could force additional clicks in the EHR or mandatory verbal scripts for pharmacists during counseling, potentially slowing throughput and increasing alert fatigue.

• Vendor & Strategic Complexity: Large health systems spanning multiple states may face a technical nightmare where a single EHR build is no longer compliant across all locations. You may need to negotiate strict indemnification clauses with vendors to ensure their "black box" algorithms meet the strictest state standards to avoid local penalties.