Artificial intelligence (AI) is no longer a future concept in pharmacovigilance (PV). It’s already reshaping how case processing gets done. From narrative writing to event coding, AI is helping teams handle increasing case volumes while maintaining consistency. But while the technology is advancing quickly, one thing remains clear in PV: human expertise is essential. In PV, human in the loop is not optional; it is a mandatory requirement.
Here are five areas where AI is already making an impact and where humans still play a critical role.
1. Case Narrative Generation: From Hours to Minutes
Writing a clear, medically meaningful case narrative is one of the most time-intensive parts of case processing. It requires structuring patient history, adverse events, and timelines into a coherent story. This becomes especially difficult during high volume periods like product launches or public health emergencies. Service providers face an added layer of complexity, since narrative format and content expectations often vary by client. AI can now generate first drafts of narratives directly from structured safety database fields. Using natural language generation, these tools transform raw data into consistent, submission-ready text in a fraction of the time.
Where humans remain essential:
AI produces the draft, but case processors still perform the final review and medical interpretation, and confirm the narrative follows the required template and regulatory expectations. This doesn’t eliminate the need for case processors; it changes their role. Instead of building narratives from scratch, they review, refine, and ensure clinical accuracy, especially for complex or high-risk cases.
2. Case Triage: Faster Prioritization, Better Focus
Triage is the first step of case processing, determining whether a case is valid, serious, expected, and reportable. Ensuring quick and robust triage allows the expert team to focus on what matters the most.
AI can automatically extract key information from the source data, such as adverse events, suspect drugs, and timelines, to determine case priority. This leads to faster classification and quicker routing of high-priority cases. AI can also flag incomplete or ambiguous reports, helping teams focus their follow-up efforts where it matters most.
Where humans remain essential:
AI takes the first pass at the case. Case processors then review that assessment and focus their attention on edge cases, ambiguous data, and judgement calls that involve regulatory nuance.
3. Causality Assessment: Decision Support, Not Replacement
Assessing whether a drug caused an adverse event is one of the most complex and subjective tasks in PV. It requires an expert medical eye and strong pattern recognition skills.
AI is great at identifying patterns across historical cases, identifying risk factors, and highlighting relevant relationships. It can provide probability scores or suggest potential causality levels.
But causality is not just pattern recognition; it is interpretation. Clinical context, confounding factors, and mechanistic understanding all need to be weighed.
Where humans remain essential:
AI highlights patterns and suggests an assessment, but medical doctors, drawing on years of clinical experience, should still hold the final causality decision, particularly where clinical nuance, uncertainty, or competing explanations are involved.
4. Data Entry from Source Documents: From Manual Extraction to Smart Capture
Case processing does not always start with a clean, fully completed AE form. It often begins with messy source documents, e.g. emails, PDFs, scanned forms, even handwritten notes.
AI-powered optical character recognition (OCR) technology and document processing tools can now extract and structure this information, even from low-quality or handwritten inputs. This shifts data entry from manual transcription to automated population of safety databases. Some systems can even identify missing data and generate follow-up queries automatically.
Where humans remain essential:
It is tempting to let the system extract everything and move straight to the next case. The real value comes from case processors assessing and approving the AI’s extraction and classification before it moves forward. Humans must still validate extracted data accuracy, resolve discrepancies, and interpret context-dependent information.
5. MedDRA Coding: Smarter, Faster, More Consistent
Coding adverse events from verbatim to MedDRA Preferred Terms is critical but may depend on individual expertise, particularly when there is a wide range of reported terms.
AI can analyze verbatim terms and suggest or assign the most appropriate PTs, even recognizing synonyms or misspellings. This improves both speed and consistency, particularly in high-volume environments. The key is for AI to use certainty scores to flag ambiguous scenarios for review, which helps maintain quality instead of sacrificing it for speed.
Where humans remain essential:
The draft coding from AI still needs systematic review by the case processor. This remains an area requiring tight control, given how central the preferred term selection is to data quality. Case processors should resolve ambiguous coding decisions and ensure that context is correctly captured in complex scenarios.
The Bigger Picture: The “Human-in-the-Loop” Model
Across all five use cases, a clear pattern emerges: AI is not replacing pharmacovigilance professionals; it is augmenting them. AI handles the high-volume, repetitive, and structured aspects of case processing. Humans step in where judgment, clinical reasoning, and accountability are required.
This “human-in-the-loop” (or “centaur”) model is likely to define the future of pharmacovigilance: combining AI’s speed with human expertise.
Final Thought
AI is already transforming case processing in practical, measurable ways. The question is no longer whether to adopt it, but how to integrate it responsibly.
The most successful teams will not be those that automate everything, but those that understand where automation adds value, and where human insight remains indispensable.
About UBC
United BioSource LLC (UBC) is the leading provider of evidence development solutions with expertise in uniting evidence and access. UBC helps biopharma mitigate risk, address product hurdles, and demonstrate safety, efficacy, and value under real-world conditions. UBC leads the market in providing integrated, comprehensive clinical, safety, and commercialization services and is uniquely positioned to seamlessly integrate best-in-class services throughout the lifecycle of a product.
About the Authors
Dobrochna Dolicka, Safety Scientist, Global Safety Writing
Dobrochna Dolicka, PhD, serves as a Safety Scientist on UBC’s Global Safety Writing Team. For the last 3 years, she has been responsible for authoring periodic safety reports such as DSURs, PBRERs, and PADERs, as well as conducting other signal management activities. Ms. Dolicka is also working on the development and implementation of automation and artificial intelligence across UBC’s comprehensive pharmacovigilance services. She holds a PhD in Biomedical Sciences as well as a master’s in molecular biology.
Christopher Henry, Safety Scientist, Global Safety Writing
Christopher Henry, PhD, serves as a Safety Scientist on UBC’s Global Safety Writing Team. With a strong biomedical and AI background, Mr. Henry has brought his unique skillset to UBC’s pharmacovigilance team. He is responsible for authoring periodic safety reports such as DSURs, PBRERs, and PADERs, as well as conducting other signal management activities for pharmaceutical products that are still in development and products that are already marketed. Mr. Henry has been leading the development and implementation of artificial intelligence across UBC’s comprehensive pharmacovigilance services. He holds a PhD in Cell Physiology as well as a master’s in health biology, Genetics, Epigenetics, & Cell Fate Control. He has worked at UBC for the past 3 years.
