European guidance on AI draws from EMA’s 2024 reflection paper and aims to promote safe, responsible AI use in healthcare, with an eye toward global convergence.

Guideline 1 centers on human-centric design and ethical alignment for AI, without detailing specific values.

Industry activity reflects momentum, including AstraZeneca’s acquisition of Modella AI to speed oncology research and Nvidia and Eli Lilly’s joint $1 billion, five-year AI lab plan in the Bay Area.

The FDA, in collaboration with the EU, released 10 guiding principles for using artificial intelligence in drug development to safeguard safety, efficacy, and quality as AI tools proliferate through the drug product life cycle.

Guideline 9 addresses life cycle management with risk-based quality oversight, issue assessment, and data drift testing during the lifetime of AI tools.

Guideline 8 requires risk-based performance assessment, including evaluation of human-AI interactions and ongoing monitoring of performance and safety with validation through testing.

Guideline 6 emphasizes data governance and documentation of data sourcing, processing, and analytical decisions while protecting privacy.

Ethics guide these efforts, with a goal of global convergence on AI topics to enable responsible innovation in collaboration with international public health partners.

Guideline 5 calls for multidisciplinary expertise to oversee AI throughout its life cycle.

The initiative highlights broader AI use across the medicines lifecycle and regulatory decision-making, while permitting controlled testing of innovative AI-driven methods.

Guideline 3 mandates adherence to applicable legal, technical, and regulatory standards for AI technologies.

Guideline 7 promotes transparent, reliable model design and development practices that enhance patient safety through sound software engineering and system design.