Pharma marketers flag high risks in AI compliance tools

A survey from Klick Health and Momentum Events reports that most pharmaceutical marketing and promotional review professionals in the U.S. lack confidence in AI for preparing regulatory compliance submissions.

Sixty-five percent of respondents said they do not trust AI for these high-stakes tasks, citing risks such as hallucinations, poor traceability and limited transparency.

Concerns identified in the study include hallucinations at 40 percent, lack of traceability or audit trail at 20 percent and limited explainability at 12.5 percent. The results show a significant gap between growing experimentation with AI tools and willingness to rely on them for regulated processes.

Demand for transparent, accountable AI systems

Alfred Whitehead, EVP, Applied Sciences, Klick: "These findings reinforce the critical need for healthcare marketing compliance tools that provide both visibility and accountability in their decision-making. The good news is that there’s an alternative to black-box AI systems, which give answers without explaining how they reach a conclusion and make it difficult to validate responses."

Julie Turnbull, SVP, Science + Regulatory, noted the emergence of glass-box AI systems designed with medical and regulatory expertise. Turnbull: "There are glass-box AI tools, built with expert human medical and regulatory decision intelligence that offer users both visibility into their workflows and reasoning behind their outputs. They are well-suited for high-risk, sensitive tasks like healthcare marketing regulatory submissions."

Klick highlights Guardrail as a glass-box example

Turnbull pointed to Klick Guardrail, described by the agency as an AI system that applies complex reasoning and a decision engine that shows its logic. The tool is built for healthcare marketing and provides traceability to support compliance while navigating regulatory frameworks. It is currently in pilot programmes with several pharma and biotech companies and is used internally at the agency for asset creation, review and MLR workflows.

BJ Jones, Chief Commercial Officer at NewAmsterdam Pharma: "The Klick Guardrail system stands out because it makes responsible AI not just possible, but practical. It was heartening to see such a thoughtful approach to ensuring safety and transparency built right into the foundation of how these tools are used to support accountability, auditing, and trust. I'm really excited to see Guardrail tackle this challenge in a way that respects reviewers and content authors alike."

Additional findings from the Klick and Momentum survey

Half of respondents said they trust AI for reviewing materials, suggesting lower comfort levels with using it for regulatory submissions. More than one-third, 37.5 percent, reported that their organisations review over one hundred assets per business quarter, while 15 percent review fewer than twenty-five.

Fifty-five percent of participants said their companies are exploring AI for review processes, and one-quarter reported they are in pilot phases. Only five percent have partially deployed AI, 2.5 percent have fully deployed it and 12.5 percent are not considering AI for this purpose.

Workforce roles and wider industry context

Respondents represented a mix of functions: 35 percent worked in MLR operations or management; 30 percent in regulatory; 15 percent in marketing; 12.5 percent in medical or medical communications; five percent in legal; and 2.5 percent in technology.

The report aligns with broader sentiment captured in the 2025 Trust in AI survey by KPMG and the University of Melbourne, which found rising concern about AI tools. Worry increased from 49 percent in 2022 to 62 percent in 2024, while 56 percent of respondents said AI use had led to errors in their work due to incorrect or hallucinated content or misinterpreted responses.

www.klick.com

Follow Marketing Report on LinkedIn