May 2025 • PharmaTimes Magazine • 12-13

// AI HEALTH //


Shape shifters

The future of regulatory submissions – how AI is transforming efficiency and compliance

Image

Regulatory submissions are more complex and resource-intensive than ever. For companies bringing new therapies and medical innovations to market, the regulatory process can be a major bottleneck.

It is not a one-time problem to solve but a constantly morphing puzzle with evolving guidelines, regional differences and exponentially growing volumes of data.

However, artificial intelligence is emerging as a powerful tool to help transform how regulatory submissions are built, reviewed and, most importantly, maintained.

While AI’s role in drug development and clinical trials is well known within the industry, its application to regulatory submissions is now becoming something that is truly exciting.

Around the world, regulators and industry stakeholders are exploring how AI can improve submission workflows, reduce manual errors and accelerate time to market – all without compromising compliance or patient safety.

AI-enabled future

The growing interest in AI within regulatory operations is not confined to a single nation or region. Major regulatory authorities are actively laying the groundwork to support AI-driven processes:

  • United States: The Food and Drug Administration has recently issued draft guidance that proposes a framework for AI’s role in drug development and regulatory decision-making. The guidance recognises the increasing number of submissions containing AI elements and promotes early interaction between sponsors and the agency to help ensure AI-generated evidence meets regulatory standards
  • European Union: The European Medicines Agency has introduced a workplan extending through 2028 aimed at leveraging AI benefits while managing potential risks. The plan explores AI’s role in automating regulatory processes, augmenting data analysis and guiding decision-making throughout the entire medicine life cycle
  • United Kingdom: The Medicines and Healthcare products Regulatory Agency has outlined an AI strategy emphasising safety, transparency and fairness. The agency plans to evolve its submission review processes to support increasing AI use in regulatory activities
  • Japan: The Pharmaceuticals and Medical Devices Agency has created sub-committees dedicated to how emerging technologies, and in particular AI, can enhance regulatory review efficiency
  • China: The Cyberspace Administration of China has introduced comprehensive regulations governing AI use, with a focus on transparent algorithms and national security. These rules require companies to disclose how their algorithms function and ensure they align with government standards, reflecting China’s broader efforts to maintain oversight of emerging technologies in healthcare and beyond.

Despite different approaches, the message from regulators is clear: AI can improve the regulatory process but must be utilised responsibly, transparently and fully compliant with regulatory standards.

Reshape submissions

The regulatory submission process is inherently time-consuming and detail-oriented. Dossier preparation requires meticulous compilation of clinical data, thorough documentation of manufacturing processes and cross-referencing thousands of pages to meet agency guidelines.

This is where AI can deliver real value.

AI technologies, particularly natural language processing and machine learning, can automate routine tasks, such as compiling submission content, extracting relevant data from trial documents and even drafting initial sections of dossiers. By automating these mundane aspects of submission preparation, AI frees up time and effort for regulatory teams to focus on higher-value activities.

Beyond efficiency, AI also enhances quality and accuracy. Algorithms can conduct automated quality checks across documents, flag inconsistencies and identify missing information, which reduces the likelihood of mistakes caused by human error.

AI isn’t just streamlining operations – it’s also strengthening regulatory compliance. When properly trained and validated, AI can help ensure that submission documents align with current agency expectations, reducing the risk of application delays or rejections.

In some cases, AI-powered systems can even monitor real-time regulatory intelligence and alert teams when guidelines change, allowing them to adjust templates or content before the submission.

Together, these capabilities can enable faster, more accurate and more compliant submissions, ultimately speeding time to market.

Caution is critical

With all of these outstanding benefits, why isn’t AI universally implemented in regulatory submissions today? It’s because the road to widespread AI adoption comes with some high hurdles:

  • Transparency concerns: Many AI models, particularly machine learning systems, are often described as ‘black boxes’, meaning it’s difficult to understand how they arrive at certain outputs. For regulators, this lack of explainability is a red flag. Agencies require clear, documented evidence to support regulatory decisions and any AI-driven content or analysis must be traceable and justifiable
  • Data integrity risks: AI systems are only as reliable as the data they are trained on. Poor-quality, incomplete or biased data can lead to inaccurate outputs – potentially delaying approvals or, worse, compromising patient safety. Ensuring data accuracy, relevance and consistency is essential before implementing AI in submission workflows
  • Compliance challenges: Introducing AI into a highly regulated, GxP environment comes with risks. AI-generated submission content must meet the same standards as manually prepared documents. That means companies must validate their AI tools, ensure robust human oversight and maintain detailed audit trails to demonstrate compliance with agency requirements
  • Risk of over-reliance: While AI can streamline certain processes, regulatory submissions still require human expertise and judgement. Agencies expect that subject-matter experts, not algorithms, will be accountable for the submission’s final content and the decision-making that goes into it.

Future-proof

The future of regulatory submissions is not about replacing human expertise with algorithms. Rather, it’s about leveraging AI as a partner to enhance decision-making and drive better outcomes.

Industry leaders and regulators alike are approaching this shift thoughtfully and, importantly, together. Many agencies are working to amend their frameworks to provide clearer guidance for using AI responsibly in regulatory submissions.

Regulatory agencies from the US, Europe and Asia have all signalled that AI-driven submissions will be welcome, provided they uphold data integrity, transparency and compliance.

At the same time, companies preparing to implement AI should take a measured, strategic approach. This includes:

  • Establishing strong data governance practices to ensure AI systems are trained on high-quality, relevant data
  • Validating AI tools to demonstrate reliability and regulatory compliance
  • Maintaining human oversight to review and approve AI-generated outputs
  • Monitoring regulatory developments to stay aligned with evolving agency expectations.

Ultimately, AI has the potential to improve submission efficiency, reduce errors and enhance compliance. However, its success will depend on responsible implementation, ongoing collaboration and a commitment to maintaining regulatory standards.

As we look ahead, AI is poised to become an integral part of regulatory submissions – not as a shortcut, but as a tool that empowers regulatory teams to work smarter, faster and with greater confidence.


Sadia Ahmed is Director, Strategic Operations and Project Management at IQVIA

0