Staying Ahead of Regulatory Revisions
In light of the evolving therapeutic landscape, increasing use of advanced technologies, and political shifts, regulatory updates are inevitable, meaning that bio/pharma companies will need to adjust quickly to new requirements.
To successfully develop and market a drug product, bio/pharmaceutical companies must adhere to strict regulations and guidelines, which have been set out to ensure quality, efficacy, and safety standards are met. However, in such a dynamic industry, approaches and technologies are continuously evolving and progressing, necessitating updates to the relevant regulations and guidelines that steer companies through the development and manufacturing lifecycle.
Recent Regulatory Revisions
The regulatory landscapes across the U.S. and European Union (EU) have undergone momentous change over recent years. “In the U.S., the regulatory landscape for the biopharmaceutical sector has undergone significant transformation with the passage of the Inflation Reduction Act (IRA) in 2022. This pivotal legislation has empowered Medicare to directly negotiate drug prices, a shift that introduces new dynamics in pricing strategies and impacts revenue streams for pharmaceutical companies,” remarks Renato Rjavec, VP Regulatory Product Management, ArisGlobal.
“Alongside these measures, the FDA has released draft guidance in 2025 for the application of artificial intelligence (AI) in drug and biologic product decision-making, emphasizing transparency, high data quality, and ongoing monitoring,” Rjavec continues. “Coupled with growing federal initiatives to fortify domestic biomanufacturing capabilities, these changes collectively signal a robust move toward increased oversight and strategic investment in the U.S. life sciences sector.”
For Frits Stulp, Partner, Life Sciences, Implement Consulting Group, a combination of three regulatory developments will considerably affect the landscape in the future. Firstly, the slow-going implementation of the ISO Identification of Medicinal Products (IDMP) standards, which is leading to a more data-driven way of working for marketing authorization holders (MAHs), he notes.
“Besides [ISO IDMP implementation], recent developments in cloud-based regulatory working (i.e., Accumulus (1) and precisionFDA (2) Trusted Regulatory Spaces) are impacting the future heavily — although it is more a development instead of a regulation,” Stulp says. “The EU Act on AI is naturally important in the way life sciences applies AI, also within regulatory affairs.”
Focusing on regulatory operations within the EU, Remco Munnik, Regulatory Information Management Expert, Arcana Life Science Consulting, points to the changes shaping product data management. These changes in the EU are being driven by “recent recommendations from the Network Data Steering Group (NDSG) and the Regulatory Optimisation Group (ROG). The objective is to establish the Product Management Service (PMS) as the trusted central source of product data in Europe, replacing the legacy system: xEVMPD,” he explains (3).
“In this plan, National Competent Authorities (NCAs) will need to align with PMS content,” Munnik specifies. “Industry stakeholders are calling for full capability to correct and enrich PMS data directly and fully support the ROG initiative, which enables [the submission of] variations directly as data changes in PMS, without formal Type IA submissions. This shift aims to streamline regulatory processes and improve data accuracy across the EU’s medicinal product landscape.”
Additionally, Rjavec highlights the bloc’s current overhaul of its pharmaceutical legislation — a move that is aimed at stimulating innovation and resolving persistent issues around drug shortages. “Central to these reforms are initiatives designed to enhance patient access, strengthen supply chain security, and introduce rigorous new regulations for AI and medical devices in the healthcare sector,” he says. “The Pharmaceutical Strategy for Europe underpins these actions, reflecting the EU’s commitment to modernize its health framework in response to evolving scientific, economic, and public health needs.”
Legislative Reform in Europe
“The EU is moving forward with major pharma law changes, now in trilogue talks,” confirms Munnik. “Key updates include shorter exclusivity periods, stricter supply requirements, and new rewards for developing antimicrobials. Generic competition should speed up thanks to broader exemptions. These reforms aim to improve access and innovation, but companies will need to adjust quickly to new regulatory demands. As an impact for regulatory operations teams, we hope to see an increase in automation with cross-functional data integration and faster turnaround expectations in future.”
Stulp reveals that the legislative reforms underway in the EU are being strongly influenced by the region’s overarching policies on sustainability, innovation potential, and technical advancement. “The pharmaceutical reforms focus on areas like supply chain resilience but also fair access to medicines for all member states,” he says.
“A particular example is the approach taken towards the management of drug shortages, where the European Shortage Monitoring Portal has gone live, using both referential medicinal product data as collected by the EMA and additional supply information provided by the MAHs,” he confirms. “This [approach] again highlights the increased focus on the role of data within the regulatory network, thereby also setting the bar for product lifecycle management by the industry, using high quality data.”
However, from a pharmacovigilance perspective, the legislation has been in place for some time and is already very descriptive for most activities, asserts Nicole Baker, CEO, Biologit. “But there is still room for harmonization and integration within internal and external systems,” she states.
Global Harmonization and Alignment
“Turning to the United Kingdom, the post-Brexit regulatory environment is characterized by efforts to harmonize with international standards while carving out an independent approach tailored to local priorities,” emphasizes Rjavec. “The U.K. is actively developing its own regulatory pathways, particularly with respect to medical devices, where enhanced post-market surveillance is a key focus. These changes position the U.K. as a market responsive to global trends, yet agile enough to implement bespoke reforms that ensure patient safety, promote innovation, and sustain competitiveness in the rapidly advancing pharmaceutical landscape.”
In efforts to align with international standards, China has undergone a rapid transformation of its pharmaceutical regulatory system over the past five years, Rjavec adds. “The National Medical Products Administration (NMPA) streamlined approvals, slashing typical review times for new drugs from years to just months, and introduced robust intellectual property protections, including patent term extensions and regulatory data protection. Government policies prioritized biotech growth, using pricing negotiations and volume-based procurement to expand patient access while containing costs,” he specifies.
“As a result, China’s drug approvals soared, its share of the global biopharma pipeline increased, and its scientific capabilities advanced, making China a key originator of innovative therapies,” Rjavec stresses. “These reforms have positioned China as a rapidly growing and increasingly influential player in the global pharmaceutical landscape, balancing affordability with a strong push for innovation.”
For other countries and regions, the implementation of electronic common technical document (eCTD) v4.0 has been a priority, reveals John Cogan, COO, Qinecsa. “[The eCTD4.0 is] a regulatory electronic submission standard which has profound implications on the level and structure of submission content and metadata,” he says. “This [focus] in turn will be very impactful for CDMOs and internal manufacturing as the chemistry, manufacturing, and control (CMC) submission components are significantly impacted by eCTD4.0.”
Guidance on AI
So far, regulatory guidance on AI has been general and foundational, but has provided a good first step for industry, reveals Cogan. “The U.S. guidance is very focused on extensive documentation — models, designs, risks, bias, validation, limitations and so on — whereas the EU is layering their guidance on top of existing GxP,” he notes. “Having these guidelines is super helpful as we identify the real, commercially viable use cases and eliminate the hot air from the AI chatter as we move from PoCs [proof of concepts] into production.”
However, those working within the bio/pharmaceutical sector should be anticipating “increasingly defined and rigorous regulatory landscape as both the U.S. and the EU introduce targeted frameworks for AI,” confirms Rjavec.
While both regulatory authorities from the U.S. and the EU are in favor of AI, there are some important provisions, stipulates Michelle Bridenbaker, COO, Unbiased Science. “Both the FDA and EMA appreciate and welcome the benefits of AI use in expediting critical regulatory processes, to help bring important medicines to market and into the hands of patients more swiftly — yet without compromising patient safety,” she says.
“The FDA updated its guidance for sponsors in January of this year [2025] (4), around using AI to produce information or data in support of regulatory decision-making around the safety, effectiveness, or quality of drugs,” Bridenbaker continues. “Its proposed framework introduces a risk-based credibility assessment to evaluate AI models used in drug lifecycle processes.”
This voluntary framework and guidance approach taken by the U.S. differs to the route being taken by the EU, which is also seeking to provide clarity around the use of AI, emphasizes Preeya Beczek, Regulatory Affairs and Compliance Expert — Beczek.COM. “The EU’s AI Act introduces binding, risk-based regulations, particularly for high-risk applications in sectors like healthcare and law enforcement. It sets out strict requirements for data governance, transparency, and mandatory conformity assessments,” she explains.
“High-risk systems — common in the pharmaceutical domain where patient safety and data integrity are paramount — are subject to enhanced requirements, including robust quality assurance, comprehensive risk management strategies, and clear transparency obligations,” concurs Rjavec.” By meeting these criteria, companies not only assure compliance but also demonstrate their commitment to fundamental rights and patient safety, which are central to the EU framework.”
Additionally, the EU’s AI Act, which is supported with EMA guidance on safe AI use across the product lifecycle, will also incorporate AI systems considered to be high risk that are used in clinical trials and regulatory processes, Munnik asserts. “In practice, AI and GenAI are increasingly used in EU regulatory operations to speed up and improve accuracy. Companies apply these tools to automate dossier creation, validate product data, and analyze large volumes of regulatory documents,” he says. “EMA supports this shift with guidance, and upcoming AI regulations will formalize standards. Early adoption helps firms stay compliant and competitive.”
To help companies and other industry stakeholders navigate the new landscape, EMA has published a reflection paper (5). In this paper, the agency notes that “applicants and stakeholders should ensure all AI/ML [machine learning] systems used in the medicinal product lifecycle are developed, validated, monitored, and documented in compliance with EU legal, ethical, and technical standards, with proactive risk management, regulatory engagement, and emphasis on transparency, data protection, and trustworthy, human-centric AI,” Rjavec remarks.
With Great Potential Comes Great Responsibility
“As can be seen in the many publications and presentations, AI has great potential to change the way of working within the industry,” affirms Stulp. “This [tool] can be applied in increasing efficiency in the administrative burden (like within regulatory and pharmacovigilance processes) but clearly also in the actual development of new medicines, by optimizing clinical study design, predictive analysis, big data set analysis and even in silico trial developments.”
Yet, according to a recent report from MIT (6), 95% of GenAI pilots fail, reveals Cogan. “On [the one hand] this shows a fail-fast mentality but on the other hand that ‘having something to say in AI’ has resulted in a lot of wasted dollars. There is no doubt there are excellent pharma AI use cases and hopefully the agencies will continue to update the guidance as more experience is gained,” he adds.
“One of the primary challenges of AI in regulated and scientific fields is the ‘black box problem’, in other words the inability of some AI systems (e.g., generic AI systems) to provide transparency in their decision-making processes. This is known as explainable AI, or XAI,” highlights Bridenbaker. “In life sciences, there must be detailed explanations for each output, showing the source of the data, the reasoning used, and any regulatory implications. This [practice] will cultivate trust and ensure that the system’s outputs can be defended in audits or reviews. For instance, in pharmaceutical research or medical diagnostics, the system must not only provide a response but also explain the source of the data, the reasoning applied, and the reliability of the outcome.”
When designing AI systems, it is key that they are capable of breaking down complex processes into smaller, manageable steps, Bridenbaker notes. “For example, a chatbot assisting with drug approval might guide the user through each regulatory stage, explaining the data requirements, ethical considerations, and compliance checks at each point,” she says.
However, the criticality of having appropriate human controls must not be forgotten, Bridenbaker stresses. “While AI can process large datasets and suggest options, humans must oversee final decisions where regulatory accountability and nuanced judgment are critical,” she asserts. “Users also need to be fully trained in using AI-enabled systems appropriately and responsibly, and in a way that mitigates risk.”
Additionally, and equally crucially, the data used to train up an AI system for the pharma industry must be domain specific and of high quality to ensure it meets regulatory standards, Bridenbaker points out. “Continuous learning models, updated with new regulatory guidelines and scientific advancements, will ensure that the system stays relevant and reliable,” she states.
“As long as the pharma industry takes on board the key principles of responsible AI use, as relevant to this strictly regulated environment, there are extensive benefits to be had, from improved targeting, curation, and accuracy of information to major operational efficiency gains,” specifies Bridenbaker.
AI systems designed for use within the bio/pharmaceutical sector should be built with potential future regulations in mind and to the highest standards available at the time, confirms Baker. It is also critical that industry stakeholders keep abreast of any new regulations or legislations about AI and other innovative approaches in a timely fashion during this period of change and development, she summarizes.
References
Accumulus Synergy. AccumulusSynergy.org, accessed Sep. 10, 2025.
precisionFDA. precision.FDA.gov, accessed Sep. 10, 2025.
EMA. Substance and Product Data Management Services. EMA.europa.eu, accessed Sep. 10, 2025.
FDA. Considerations for the Use of Artificial Intelligence to Support Regulatory Decision-Making for Drug and Biological Products. Draft Guidance Document, January 2025.
EMA. The Use of Artificial Intelligence (AI) in the Medicinal Product Lifecycle. EMA.europa.eu, accessed Sep. 10, 2025.
MIT NANDA. The GenAI Divide: State of AI in Business 2025. Report, August 2025.