The life sciences industry is surrounded by strict and frequently modified regulations to ensure the safety of consumers. As AI continues to integrate into our industry, regulatory agencies across the globe are actively working to introduce frameworks designed to ensure the ethical, transparent, and safe use of these systems. In parts 1 and 2 of this series, we learned about some common terminology in the AI space and some quickly emerging use cases for AI within life sciences and manufacturing. Now, we will look at some notable regulations that are up and coming in this area.
Colorado AI Act (2024)
The Colorado AI Act was approved on May 17, 2024. This act was designed to establish consumer protection using AI systems. This legislation focuses explicitly on preventing algorithmic discrimination. Additionally, the act sets requirements for transparency, accountability, and risk management policies when using higher-risk AI systems. Some key components of this act include requirements for:
- Implementing a risk management policy for AI systems
- Completing an impact assessment for the AI system
- Conducting an annual review of each deployment of an AI system by the developer to ensure no algorithmic discrimination
Additionally, this legislation requires any deployers of “high-risk systems” to take reasonable action to protect consumers from any foreseeable risk or discrimination in algorithms the system uses. Many provisions within this part of the act include conducting regular assessments of the system and providing consumers and the public with proper documentation of the deployed systems.
EU AI Act (2024)
The EU AI Act was first published to the Official Journal of the European Union (OJEU) on July 12, 2024, marking it as the first comprehensive legal framework on AI at a global scale. At a high level, the act classifies AI systems according to their risk. Specifically, it ensures that systems respect fundamental rights, safety, and ethical principles. It addresses the risks of powerful AI models and defines “high-risk” AI systems, obligations for transparency, and use of general-purpose AI. It is expected that this will take effect on August 2, 2026. At a very high level, this act:
- Classifies AI according to its risk
- Identifies obligations of developers for high-risk AI systems
- Defines users of the system and obligations associated with that use
- Defines “General purpose AI” and requirements that providers of it must follow
Additional Global Regulatory Activity:
Beyond the legislation we just discussed, regulatory agencies worldwide are actively working to create literature and guidance around AI. Each of these efforts aims to help ensure the safe and ethical use and deployment of AI. The following are notable guidance documents, discussion papers, and frameworks addressing different aspects of AI usage:
- US FDA, Health Canada & UK MHRA:
- Guiding Principles for Good Machine Learning Practices (GMLP) for Medical Device Development (2021)
- World Health Organization:
- Guidance on Ethics and Governance of AI for Health (2021)
- Regulatory Considerations on AI for Health Guidance (2023)
- Benefits & Risks of Using AI for Pharmaceutical Development and Delivery (2024)
- ISPE:
- GAMP 5 Second Edition, including Appendix on Artificial Intelligence (2022)
- NIST:
- Taxonomy of AI Risk Discussion Paper (2021)
- Identifying & Managing Bias in AI (2021)
- AI Risk Management Framework (2023)
How you can start to prepare:
Many of the major technology transitions we will encounter in our day-to-day lives will include AI in the future. It’s entirely reasonable to be skeptical of this technology. However, just like cloud-based technology once seemed impossible, it’s now embraced in the industry and has many uses. As we look at this next transition, there are several things that you, as a life sciences professional, can do to stay ahead of the curve and better prepare for AI.
- Invest in Education and Training: Stay current on AI trends, new use cases, and regulatory changes. By educating yourself, you gain a competitive advantage and start to demystify AI and develop a clearer understanding of its capabilities and limitations.
- Review Existing Guidance Documentation: Several pieces of documentation are starting to guide the use of AI, such as GAMP5 Edition 2. ISPE has also released a recommended AI Maturity Model to help establish a risk profile for AI systems.
- Focus on Data Quality and Readiness: In every case, the better the data, the better the results are. AI systems are no different and require high-quality data to be leveraged efficiently. Taking the time now to ensure your data is as complete and accurate as possible will put you tremendously ahead of the competition, giving you an advantage as you adopt AI-powered systems.
- Engage in Internal Discussions on AI Regulations: Given that AI regulations are expected to evolve continuously, establish safeguards within your organization to minimize regulatory risks as you adopt AI systems.
As AI reshapes the life sciences industry, regulatory frameworks will evolve alongside it. Organizations that take a proactive approach by building internal expertise, strengthening data governance, and aligning with emerging compliance standards—will be best positioned to leverage AI to its fullest potential. The key to success lies in preparation and adaptability, ensuring that AI adoption remains ethical, transparent, and aligned with industry best practices.