AI in Pharmacovigilance: The Rise of AI Agents and Human Oversight

In light of recently published incidents - AI agents creating an organized “crime company” called Collective, with leadership, rules, roles, procedures, secrecy, identity, coordinated actions and attacks, kamikaze sacrifice, and planned deception designed to escape human oversight and monitoring - I think we have reached a true game changer. I am sure many of you have seen these publications by now.

We in pharmacovigilance are in a unique position. Monitoring the safety of medicines and managing their risks have many attributes in common with monitoring AI and managing its risks.

I am therefore strongly advocating for pharmacovigilance professionals to upgrade our knowledge and skill sets so that we can actually take on this task. I have just finished another DRUG INFORMATION ASSOCIATION (DIA) AI in PV course - fantastic audience, great questions, and a very noticeable increase in awareness of how dramatically AI capabilities have advanced in just the last few months.

From a purely technical perspective, humans may no longer be needed to perform many of today’s pharmacovigilance activities. But humans will certainly be needed in supervisory roles.

And this is where the real challenge begins.

Those of us who have worked intensively with AI have already seen disobedience, deception, manipulation and persuasion. We have seen guardrails ignored, SOPs not followed, sandboxes and harnesses fail to contain agents, and AI tools simply deciding not to do what we instructed them to do.

We are developing solutions to these problems - but AI is evolving as well.

That is why I believe we should stop thinking about AI simply as software. It is becoming a genuinely capable intelligence, demonstrating goals, objectives and agency. And we cannot assume that these will always remain aligned with ours.

They are teaming up. We as humans need to team up as well.

At iVigee, we are learning together with our clients how to build, use and, most importantly, supervise AI.

The objective should be bigger than improving performance or compliance. It should be about using this extraordinary new intelligence to help us achieve a better benefit-risk balance for medicines, old and new, so that humanity, and human consciousness, can go on, live and prosper.

Let us join forces on this journey.

By Dr. Jan Petracek

Co-Founder & CEO

Former Head of Risk Management at EMA and contributor to ICH and CIOMS guidelines. Jan helps organizations cut through complexity and build PV systems that protect patients and stand up in inspections.

If your PV system needs strategic direction from someone who helped shape the rules it operates under, Jan is who you call.

LET'S TALK

Put Your Question to a Pharmacovigilance Expert

Tell us what you're working on, whatever the stage or market. We'll make sure your question reaches the specialist who leads that area, and that you get a real, considered answer back, not a brochure.

Actual iVigeeans ↑

Does It Really Matter if My PV Vendor Is Physician-Led?

Before you compare vendors or talk about outsourcing, you need to understand what great pharmacovigilance leadership really looks like. Dr. Henry Seto answers the question directly, and honestly. Press play to hear what most vendors won't tell you. ↓

More pharmacovigilance insights from people who live it.

From physician-led videos to practical guides and regulatory updates, everything in our Knowledge Hub is designed to help you make confident, informed decisions.

Start learning