This week, I had the pleasure of attending the DPHARM conference in Philadelphia, where innovative ideas in clinical development were the focal point of every conversation. Among the various engaging sessions, one that stood out to me was "How Pharma is Adopting GenAI Across Clinical Development." The session, led by industry leaders such as Dave Apfel from Johnson & Johnson and Henry Wei from Regeneron, provided critical insights into how generative AI is transforming clinical trials and broader pharmaceutical operations.

As someone who strongly believes that artificial intelligence exists to expand human-to-human interaction, I found this session particularly aligned with my thoughts. One of my favorite quotes that I’ve said time and time again is this: Artificial intelligence will not take your job. However, someone who knows how to leverage artificial intelligence might. This belief stems from the idea that AI is here to support and enhance human connections, not replace them.

Breaking Down Barriers with GenAI

One of the key insights from the session was how companies like Moderna are democratizing AI by making it accessible to all employees, not just those in tech or data science roles. Moderna’s creation of an AI academy, followed by hackathons and exclusive AI clubs, exemplifies their efforts to make AI a tool for every curious mind in the organization. By opening up GPT creation to all employees through OpenAI Enterprise, they now boast over 750 AI agents that help with everything from writing emails to reviewing dosage data. This is a prime example of what I would call Citizen AI—the widespread adoption of AI by individuals without a technical background.

Citizen AI vs. Generative AI: Overlapping Roles in AI’s Future

There has been much discussion around the concept of Citizen AI, which I now see as the broader democratization of AI use—making AI accessible to non-technical users across organizations and society at large. Citizen AI includes not only Generative AI but also a variety of AI tools like data analytics, automation, and predictive modeling. The key idea is that these tools allow everyday users to harness the power of AI without needing technical expertise.

Generative AI, on the other hand, is a subset of AI focused on creating new content—like text, images, or insights based on existing data. It certainly plays a role within the Citizen AI framework, as tools like GPT models enable non-technical users to generate reports, summarize complex data, or even write emails. However, Generative AI also exists outside the scope of Citizen AI, especially in advanced applications that require more technical manipulation or deeper machine learning knowledge.

Rather than thinking of Citizen AI and Generative AI as having a parent-child relationship, it's more accurate to envision them as overlapping concepts, like a Venn diagram. Citizen AI includes the more accessible forms of Generative AI, but Generative AI extends into more technical, niche areas that may not be accessible to the average user.

Getting Started is Half the Battle

As Dave Apfel of J&J emphasized during the session, one of the challenges in adopting GenAI at scale is the inherent risk-averse nature of the pharma industry. His advice was simple but profound: “You just have to get started.” Pharma companies need to embrace pilot projects, both internally and externally, to explore GenAI's potential without waiting for the perfect moment. The message was clear—while many are intrigued by the possibilities of GenAI, few have successfully scaled it.

I couldn’t agree more with this sentiment. The key to leveraging AI effectively is simply to start. It’s like using your iPhone—you didn’t have a reference manual when you first picked it up, you just started using it. The same applies to AI, especially for those adopting Citizen AI. If you find yourself needing more, whether it's deeper machine learning capabilities or advanced AI development, there are plenty of options available, from low-code solutions to full-scale AI development platforms.

The Data Dilemma

Another powerful theme from the session was the vital role of data in AI success. As Shameer Khader from Sanofi noted, while the technology is impressive, the quality and structure of the data used to train AI models are even more critical. GenAI is only as good as the data you feed it. This means that companies need to focus on developing robust data models if they want to see impactful results. In a way, it’s a reminder that AI is not a magic solution—it’s a tool that requires high-quality input to yield high-quality output.

Real Use Cases, Real Results

One of the most valuable parts of the session was the discussion around real-world applications. At Regeneron, Henry Wei shared how posing humanity-level questions to AI can separate hype from reality. As an example, try presenting AI with a math test at a 4^th grade level, then another at an 8^th and 12^th grade level. In all seriousness – try it! The idea here is that by presenting an AI model with human-like tasks, you will start to build understanding about how much trust you should – or should not put into AI’s ability to “do your work for you.”

By framing the problems GenAI is designed to solve in such terms, it becomes easier to gauge whether it’s capable of driving meaningful results. This pragmatic approach to GenAI adoption—testing its capabilities with real challenges—shows the importance of critical evaluation in the implementation of new technologies.

Closing Thoughts: Why FOMO Might Be Pharma’s Friend

The session wrapped up with an intriguing point: FOMO, or the fear of missing out, can be a significant driver for innovation. As more companies successfully implement GenAI, others will inevitably feel the pressure to keep up. This human trait, often viewed negatively, could be one of the biggest accelerators for AI adoption in pharma.

The key takeaway from this session? Don’t wait. GenAI is here, and it’s time for pharma companies to dive in, test its limits, and see how it can improve everything from clinical trials to everyday administrative tasks. As the speakers advised: Start small, stay curious, and let the results speak for themselves.

For me, this reinforces my belief that AI is a tool meant to supplement and enhance human-to-human interaction. It’s accessible to everyone, and the best path forward is to embrace Citizen AI and Generative AI. The key is to just try—whether it's in your workflows, communications, or clinical operations. If you need more advanced solutions down the road, those are available too. But the most important step is starting.