A GC I spoke with recently had been asked by her CEO to present the company’s AI governance framework at the next board meeting.
She had three weeks. She had no framework.
What she had was a collection of vendor terms she’d reviewed, a draft acceptable use policy that hadn’t been finalized, and a vague sense that something needed to happen with HIPAA before anyone ran clinical data through an AI tool.
That’s where most biopharma legal departments are right now.
Not unprepared exactly. Aware that something needs to exist — not sure what it needs to cover, what the priorities are, or how to present it to a board that is asking because they read something in the Wall Street Journal, not because they have a specific concern.
The board question is the visible problem. The regulatory one is the one that matters more.
FDA has issued guidance on AI in drug development and manufacturing. The guidance is still developing, but the direction is clear: if you’re using AI in clinical trial management, pharmacovigilance, or GMP manufacturing, you need documentation. What tools you’re using, how they’re validated, what data they’re touching, and how you’re overseeing the outputs.
HIPAA applies wherever patient health information touches an AI system. That means the BAA question — does your AI vendor have a signed Business Associate Agreement — isn’t optional. It’s the first question in any HIPAA enforcement conversation.
And acquirers are starting to ask about AI governance in M&A diligence the same way they started asking about data privacy programs five years ago. Not as a nice-to-have. As a line item.
Here’s the thing about AI governance in biopharma legal: the framework doesn’t need to be perfect to be valuable. It needs to exist, be documented, and be defensible.
A GC who can walk into a board meeting with a written AI use policy, a data classification framework, a vendor assessment process, and a clear explanation of where HIPAA applies is ahead of 80% of the industry right now. That gap won’t last. But it exists today.
The legal departments that will be positioned well in two years aren’t the ones waiting for the law to settle before they build a program. They’re the ones building the framework now and updating it as standards develop.
I built an 80-item AI governance checklist for biopharma legal departments.
It covers eight areas: AI use policy, data classification and protection, vendor and tool assessment, HIPAA and regulatory compliance, contract and IP risk, employee training and acknowledgment, incident response, and board reporting.
Each item is tagged HIGH, MEDIUM, or LOW priority. Start with the HIGH items — those are the ones a regulator, acquirer, or plaintiff is most likely to ask about. Each section includes a plain-language note on why it matters and what the gap costs you.
It’s a Word doc. Fillable, assignable to owners, shareable with your board. $127. Instant download.
One thing worth saying plainly: this checklist isn’t a legal opinion on what your AI governance program must look like. AI law is developing fast and the standards will shift. It’s a structured starting point — the 80 questions you should be able to answer before someone else asks them

