AI in Clinical Trials: What Talent Will Be Needed as Regulation Catches Up?
October 20267 min read
AI in Clinical Trials: What Talent Will Be Needed as Regulation Catches Up?
Artificial intelligence (AI) is moving deeper into clinical development.
From patient selection and data analysis to predictive modelling and trial design, its potential is significant. But as its influence on evidence generation grows, so does the need to demonstrate that AI-supported decisions are reliable, traceable, and appropriate for their intended use.
For life sciences companies, this creates a talent challenge.
The next phase of AI adoption in clinical trials will require more than machine learning engineers. Companies will increasingly need professionals who can connect AI and data science with biostatistics, clinical operations, regulatory affairs, quality, and governance.
Regulators are already moving in this direction. The U.S. Food and Drug Administration (FDA) and European Medicines Agency (EMA) published 10 joint principles for good AI practice in drug development. These cover areas including clear context of use, multidisciplinary expertise, data governance, performance assessment, and lifecycle management.
The question for employers is therefore changing from “How do we hire AI talent?” to “What combination of talent will allow us to use AI effectively, and demonstrate that we are using it responsibly?”
How is AI being used in clinical trials?
AI and machine learning can support multiple parts of clinical development.
The EMA has highlighted potential applications including selecting patients based on disease characteristics or other clinical parameters, as well as supporting data recording and analysis that may ultimately contribute to regulatory submissions.
These capabilities could change how sponsors identify suitable participants, analyze complex datasets and generate evidence. They also sit alongside wider advances in adaptive trial design, predictive modelling and data-driven drug development.
EPM Scientific is already seeing demand across clinical development for expertise spanning clinical operations, clinical data management, biostatistics and machine learning.
But applying AI to clinical research is different from deploying it in a lower-risk commercial environment. If an AI model influences evidence about a medicine's safety or effectiveness, sponsors need to understand how the model works, what data it relies on, how its performance is assessed, and where human oversight remains necessary.
That makes the people surrounding technology critical.
Regulation is catching up with clinical AI
Regulators are not waiting for AI adoption to settle before addressing these questions.
The FDA's January 2025 draft guidance proposed a risk-based credibility assessment framework for AI models producing information or data intended to support regulatory decision-making on drug safety, effectiveness, or quality.
The EMA's reflection paper similarly identifies risks around patient safety, clinical study integrity, bias and the reliability of AI and machine learning applications.
One particularly important principle is multidisciplinary expertise, because it has direct workforce implications.
AI in clinical trials cannot sit exclusively within a data science function. Sponsors will need people capable of translating between technical teams, clinical experts, statisticians, quality professionals, and regulators.
As Freddy Cruz, Vice President at EPM Scientific, explains:
As pharmaceuticals and biotechs are using AI more within their trials, to basically predict data that goes towards the trials, [regulation] is still very far away from where anything is right now.
His comments reflect a broader challenge. Technology can move faster than established processes, but clinical development cannot compromise evidence quality or patient protection simply to keep pace.
Five talent areas clinical AI teams will need
The result is likely to be growing demand for hybrid teams rather than one universal “AI in clinical trials” role.
Strong technical talent remains fundamental.
Freddy describes one of the rarest profiles in the current market as a specialist with a strong track record in building generative models. He also highlights demand for machine learning research engineers and machine learning engineers who can scale workflows, prepare data, run experiments and improve model infrastructure.
For clinical organizations, technical capability will increasingly need to be combined with an understanding of regulated evidence generation.
The hiring challenge is already considerable. Some life sciences companies are competing with major technology and AI organizations for the same machine learning talent.
AI is unlikely to replace the need for rigorous statistical expertise. Instead, it increases the importance of professionals who can determine whether new analytical methods are scientifically appropriate and sufficiently robust.
This creates an important upskilling opportunity. Rather than building every AI capability through external hiring, organizations can identify experienced biostatisticians and data professionals who can develop stronger machine learning literacy.
ICH E6(R3) also places clear emphasis on data integrity, traceability, security and fit-for-purpose computerized systems throughout the clinical trial data lifecycle.
As regulatory expectations develop, organizations will need professionals who understand both AI risk and pharmaceutical regulation.
This could increase the value of regulatory affairs, quality assurance, and governance specialists who can assess questions such as:
- What is the model's intended context of use?
- Which data was used to develop and validate it?
- How will bias and performance be monitored?
- What documentation is required?
- What happens when a model or its underlying data changes?
- Where must human oversight remain?
Governance is no longer a consideration that can be added once a model is ready for deployment. For employers, the practical implication is that governance capability should be involved early in AI projects rather than becoming a final approval step.
One of the most important profiles may not be a new AI job title at all.
Clinical scientists, physicians, clinical operations professionals and other domain specialists understand what constitutes meaningful evidence and how trials operate in practice. Their knowledge is essential when determining whether an AI output is useful rather than merely technically impressive.
Freddy sees adoption as a key challenge when technical teams and scientific users are disconnected.
Once domain experts and machine learning or engineering teams work together effectively, he says, experimental use can generate better-quality data, which can then improve the model and its outputs.
This feedback loop is one reason cross-functional capability could become a competitive advantage.
Clinical AI is only as credible as the systems and data supporting it.
That puts greater emphasis on professionals who understand data quality, lineage, privacy, validation, documentation and computerized system controls.
The FDA and EMA's good AI practice principles explicitly address data governance, documentation, model development, performance assessment and lifecycle management.
Companies therefore need to think beyond the person developing the algorithm. They need talent and processes capable of demonstrating why its outputs can be trusted.
Hiring for AI in clinical trials requires clarity
Finding these skills is only part of the problem. Employers must also convince scarce candidates to join. Freddy identifies clarity as one of the most important factors in attracting and retaining AI talent:
The clarity piece is the biggest thing, from the interview process, expectations of the role going into it, of the mission, the team.
That matters particularly for emerging clinical AI positions, where responsibilities can easily become unclear.
Before going to market, hiring teams should define the problem the person is expected to solve, where the role sits between clinical and technical functions, how success will be measured, and what authority the individual will have.
Otherwise, employers risk searching for an “AI expert” without first establishing whether they actually need a research engineer, biostatistician, data governance specialist, regulatory professional, or a multidisciplinary team.
Build multidisciplinary AI capability before regulation demands it
The direction of travel is becoming clearer. Regulators are focusing on risk, context of use, data governance, documentation, performance, and human oversight. At the same time, life sciences companies are expanding the ways AI can support clinical development.
That intersection will shape hiring. The strongest clinical AI teams are unlikely to consist solely of AI specialists. They will combine machine learning expertise with biostatistics, clinical knowledge, regulatory understanding, quality, and data governance.
For employers, this means workforce planning should start before a specific regulatory requirement creates an urgent vacancy. Assess which skills already exist internally, where professionals can be upskilled and where specialist external talent is necessary.
EPM Scientific supports pharmaceutical, biotechnology and life sciences organizations across research and development, including clinical development, clinical operations, clinical data management, biostatistics and specialist machine learning talent.
As AI changes the skills required across clinical development, access to specialist talent and current market insight can help organizations build teams around the outcomes they need rather than hiring for AI for its own sake.
Looking to strengthen your clinical, data, or AI capabilities? Speak to EPM Scientific about building the specialist talent required for the next stage of clinical development.
Find the talent your business needs
Submit your vacancy to EPM Scientific and connect with our specialist recruitment consultants. We’ll work with you to understand your hiring requirements and identify the talent your business needs to succeed.
Let's discuss your hiring needs
Looking for specialist recruitment support? Request a call back from EPM Scientific and speak with a consultant who understands your market, hiring challenges, and business priorities.