Regulatory bodies and legislatures are making moves that will impact how life sciences organizations use AI. Is your business ready?
As we’ve been writing about for some time, Artificial Intelligence (AI) is changing the life sciences paradigm across supply chain management, warehouse management, precision medicine, and other aspects of drug development.
Since 2021, the U.S. FDA has received more than 100 annual drug and biologic application submissions that include AI and Machine Learning (ML) components, covering everything from drug discovery and clinical research to post-market safety surveillance and advanced pharmaceutical manufacturing.
The world as we know it is changing fast, and pharmaceutical companies, biotech firms, and medical device manufacturers are all trying to implement the best AI solutions today while futureproofing their businesses for the new innovations of tomorrow.
It can be difficult for life sciences companies to decide when and how to embrace AI for maximum impact, especially since regulatory bodies like the FDA and the EU’s EMA have yet to put definitive stakes in the ground. To date, an FDA discussion paper and an EMA reflection paper are the primary sources of regulatory information on the topic, but they raise more questions than they answer.
The goal of both papers is to frame the discussion around different areas of consideration that are important for policy development. In the area of pharmaceutical manufacturing, for example, the FDA highlights these key issues:
These and other considerations described in the paper generally revolve around data and its integrity, quality, collection, management, storage, and analysis. Special attention is given to unintended biases that may impact outcomes, as well as issues around the explainability of AI outputs.
As recently as October 2023, the FDA—along with the U.K.’s Medicines and Healthcare products Regulatory Agency (MHRA) and Health Canada—published a paper called “Predetermined Change Control Plans for Machine Learning-Enabled Medical Devices: Guiding Principles.” With a focus on quality management, the short paper describes how predetermined change control plans that are focused, risk-based, evidence-based, and transparent can help:
New AI laws will certainly impact any official regulatory guidance given to life sciences organizations. Right now, the most sweeping AI legislation has been proposed by the European Commission and is expected to impact businesses in 2024 or 2025.
The Regulation on Artificial Intelligence will have repercussions for life sciences organizations that use AI systems to develop products for European citizens, even if the companies aren’t based in Europe. The AI Regulation complements several existing legal frameworks, including GDPR data protection and CE product safety marking.
The proposed legislation categorizes AI systems by risk. Low-risk solutions will likely be subject to basic controls that apply to all AI systems, such as informing users they’re interacting with AI and labelling content that’s artificially created or manipulated, such as “deep fakes.”
High-risk AI systems include those intended for use as a safety component of products (or which are themselves a product), such as regulated medical devices. A second high-risk category covers AI systems that impact fundamental rights. Pharma AI systems, for example, are high-risk if they process data that constitutes the “biometric identification and categorisation of natural persons.”
While the key regulatory controls for high-risk AI are expected to fall on system providers, end users of such systems are also likely to be subject to certain requirements. These include using systems as intended, ensuring that input data is relevant to the intended purpose, monitoring systems, flagging suspected risks, incidents, and malfunctions, and keeping detailed logs.
Failure to comply with the AI Regulation will likely result in substantial penalties. Proposed fines are up to €30m or 6% of global turnover, whichever is higher.
In addition to objecting to high fees, critics of the proposed legislation believe that it’s too broad in scope, touching technologies that aren’t always considered AI. The fear is that this over-inclusiveness will suppress innovation and empower broad enforcement that could hurt companies more than help them.
While the industry waits for governments to come up with policies and regulations, pharmaceutical companies, biotech organizations, and medical device manufacturers have been exploring the use of AI and its impact on medical therapies.
The IFPMA, for example, has issued a set of principles that encourage a safe and ethical environment for data use, particularly in the case of algorithmic decision-making. The principles have been offered as a starting point for considering how internal processes, controls, operations, and policies can incorporate AI in an ethical and responsible way, by:
Many individual life sciences companies have also published statements about the use of AI. The vast majority of these statements share the understanding that while AI offers tremendous potential it is also fraught with risk and limitations. To ensure success, most companies realize they need a long-term framework that adequately addresses explainability, accountability, inclusivity, outcome equality, and data security.
Human oversight of AI tools will remain a top priority, and companies using AI will be challenged to ensure that all algorithms are trained on data that meets rigorous inclusion/exclusion criteria and is as accurate and free from bias as possible.
To prepare for the day when AI dominates the life sciences industry, companies should start auditing their existing—and planned—data management, analytics, and AI-enabled systems as soon as possible.
If you haven’t undergone any digital transformation at your organization, now is the time to do so. At minimum, your organization should have a powerful, industry-compliant Enterprise Resource Planning (ERP) system that centralizes data and automates processes.
Many life sciences organizations are using outdated ERP systems.
Talk to AXP Pharma about digitalizing your company today for
greater efficiency and cost savings.
Beyond its unparalleled security, affordability, and scalability, Microsoft Dynamics 365 is a great choice for life sciences for two very important reasons:
AX for Pharma 365™ Is the Most Complete ERP Solution for Life Sciences
Convenient, configurable modules are purpose-built for life sciences organizations. These modules
can be added as needed, making it easy to scale the solution so it grows with your specific needs.
Leading pharmaceutical, biotech, and medical device companies throughout the U.S., Europe, and Australia rely on AX for Pharma to manage their critical life sciences data. To learn how we can help your business prepare for the coming AI revolution and all the accompanying regulations, please contact us today.