DIA Europe: EMA to release new AI work plan

BASEL, SWITZERLAND – The European Medicines Agency (EMA) is set to release an updated work plan outlining its activities related to artificial intelligence (AI) over the next three years. This announcement was made by Luis Pinheiro from the EMA’s data analysis and methods team at DIA Europe 2025 on 18 March.

Pinheiro discussed how AI applications are utilized by industry and regulators, focusing on AI and machine learning (ML) tools from drug discovery to the postmarketing phase.

"We are finalizing the work plan revision, but we still have a month before we complete it,” Pinheiro said.

The updated work plan will replace previous version released by EMA and the Heads of Medicines Agencies (HMA) in December 2023. (RELATED: EU agencies adopt workplan on AI in medicines regulation, Regulatory Focus 18 December 2023)

Pinheiro said the upcoming work plan will have four “swim lanes” detailing where EMA will focus its activities over the next three years. These include guidance, policy, and product support; tools and technologies; collaboration and change management; and experimentation. “These swim lanes all feed off each other,” he said.

In the experimentation area, he said the report will detail the activities of its digital innovation lab, which was established in 2021 to accelerate digital innovations. EMA also has an analytics center of excellence to explore the use of analytics, AI/ML, and robots, as new techniques for process analytics. The report will also share the results of its health data lab pilot, which develops innovative analytics to extract information from healthcare data.

He said that the future workplan will include forming a new network data steering group and assessing the implementation of the EU AI Act.

Pinheiro emphasized that a vital aspect of EMA’s work in the AI/ML field is training staff to improve their skills and proficiency in this area. Last August, the EMA released a set of guiding principles regarding the use of LLMs in regulatory science and medicines regulation. This guide aims to educate staff on effectively utilizing these tools. The principles outlined are designed to help leverage LLMs in a safe and responsible manner within their regulatory network. (RELATED: EMA, HMA publish large language model principles for regulators, Regulatory Focus 9 September 2024)

From an industry standpoint, Susan Sandler of Johnson & Johnson told the meeting that AI/ML can be leveraged across several areas to support overall drug development and regulatory efficiency.

These tools can be utilized in various non-clinical aspects of drug discovery. They can help to enhance understanding of diseases, identify new targets, predict potential risks, and select compounds. In the manufacturing sector, these tools can assist in process development, monitoring, and batch release.

In clinical trials, AI/ML tools can be used to develop novel endpoints, increase patient diversity, and inform patient selection. After a drug is authorized, these technologies enhance signal detection and automate processes and workflows. In regulatory operations, they can facilitate the interrogation of regulatory information sources, create documents, and translate materials.

In the regulatory realm, Nicolas Löffler-Perez, a data scientist at Swissmedic, described how scientists at his agency are using an AI/ML-enabled tool dubbed “Nightcrawler” to search for illegal products on the Internet. The tool was developed by the Swissmedic 4.0 innovation lab to promote interdisciplinary work and design new digital business models.

DIA Europe: EMA to release new AI work plan

Related topics

Recommended content

Welcome to the new RAPS Digital Experience