Standards, Guidance & Notices
Showing 11–18 of 18
ISO
ISO/IEC 42005:2025
Information technology — Artificial intelligence — AI system impact assessment
Specifies processes and requirements for assessing the impacts of AI systems on society, individuals and the environment. Intended for use in conjunction with ISO/IEC 42001. Serves as a framework for impact assessment in medical AI development.
Published: 2025-05-28
IMDRF
IMDRF/AIML WG/N88 FINAL:2025
Good machine learning practice for medical device development: Guiding principles
Final document of the IMDRF AI/ML-enabled Working Group setting out 10 guiding principles for Good Machine Learning Practice (GMLP), intended to promote the development of safe, effective, and high-quality AI-enabled medical devices across the total product life cycle. The principles cover: understanding the intended use and leveraging multidisciplinary expertise; good software engineering, medical device design, and security practices; clinical evaluation using datasets representative of the intended patient population; independence of training and test datasets; fit-for-purpose reference standards; model choice and design tailored to the available data and intended use; assessment focused on human-AI interactions and the performance of the human-AI team rather than the device in isolation; testing under clinically relevant conditions; clear, essential information for users; and monitoring of deployed models and management of re-training risks. The document presents the principles as a call to action for standards organizations, regulators, and other bodies to further advance GMLP, and notes that generative AI may heighten their importance.
Published: 2025-01-29
FDA
CDRH
CDRH
FDA-AI-DSF-Draft-2025
Artificial Intelligence-Enabled Device Software Functions: Lifecycle Management and Marketing Submission Recommendations
FDA draft guidance (January 2025) on lifecycle management and marketing submissions for AI/ML-enabled SaMD. It provides specific recommendations on training data management, performance monitoring, and change management (PCCP: Predetermined Change Control Plan), including recommendations on handling adaptive algorithms and on describing postmarket performance monitoring plans. Still at the draft stage, pending finalization.
Published: 2025-01-07
JFMDA
Notice
Notice
jfmda_20241226_146b6000
Guidance for Companies on the Development of AI Medical Devices Using Pseudonymously Processed Information, Version 1.0
On December 25, 2024, Team 1 of the JFMDA Subcommittee on Handling of Personal Information compiled and published the "Guidance for Companies on the Development of AI Medical Devices Using Pseudonymously Processed Information, Version 1.0." On the same date, the Medical Device Evaluation Division of the MHLW Pharmaceutical Safety Bureau issued an administrative notice providing information on the guidance. As explained in Chapter 0 (Background and Purpose) of the guidance, developing AI-enabled medical devices generally requires large volumes of medical data, such as clinical information, images (CT, MRI, etc.), and other non-text information (ECG, EEG, etc.). However, medical information held by medical institutions generally constitutes personal data under the Act on the Protection of Personal Information, and it is usually not collected for the purpose of AI medical device research and development.
Published: 2024-12-26
FDA
CDRH
CDRH
FDA-PCCP-AI-DSF-2024
Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence-Enabled Device Software Functions
FDA final guidance (December 2024) on Predetermined Change Control Plans (PCCPs), which streamline change management for AI-enabled medical devices. Modifications described in an authorized PCCP can be implemented without a new premarket submission for each change. It specifies the three elements a PCCP should include: a description of modifications, a modification protocol, and an impact assessment. It puts into practice the statutory framework of section 515C of the FD&C Act, added by FDORA.
Published: 2024-12-04
EU
Law
Law
Regulation (EU) 2024/1689
Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence and amending Regulations (EC) No 300/2008, (EU) No 167/2013, (EU) No 168/2013, (EU) 2018/858, (EU) 2018/1139 and (EU) 2019/2144 and Directives 2014/90/EU, (EU) 2016/797 and (EU) 2020/1828 (Artificial Intelligence Act)
EU regulation establishing a risk-based framework for artificial intelligence, dated 13 June 2024 and published in the Official Journal on 12 July 2024. It sets out prohibited practices and requirements for high-risk AI systems. Medical device AI becomes high-risk where it is a safety component of a device, or is itself a device, and is subject to third-party conformity assessment by a notified body (MDCG 2025-6). Requirements for high-risk systems cover data governance, technical documentation, transparency, human oversight, and accuracy, robustness and cybersecurity; a single set of technical documentation is drawn up for the MDR or IVDR and the AI Act. Article 99 sets maximum fines of EUR 35 million or 7% of worldwide annual turnover for prohibited practices, and EUR 15 million or 3% for breaches of operator obligations, whichever is higher. As amended by Regulation (EU) 2026/1744, the high-risk requirements apply from 2 December 2027 for Annex III systems and from 2 August 2028 for Annex I products, which include medical devices.
Published: 2024-07-12
EU
MDCG
MDCG
MDCG 2023-4
Medical Device Software (MDSW) – Hardware combinations Guidance on MDSW intended to work in combination with hardware or hardware components
MDCG 2023-4 addresses MDSW that cannot achieve its intended medical purpose without hardware or a hardware component, setting out three regulatory scenarios (desktop PCs and cloud platforms are out of scope): (1) the hardware qualifies as an accessory to the MDSW under MDR Annex VIII classification rules; (2) the hardware is placed on the market as a standalone medical device -- as part of a system (Art. 22), a combination (Art. 2(1)), or an integral component -- in which case the manufacturer may rely on the hardware's MDR conformity for general safety and performance requirements, though clinical evaluation must still cover the combined intended purpose; (3) the hardware is part of a general consumer or wearable product with no medical device/accessory status, in which case the MDSW manufacturer bears sole responsibility for the safety, performance and reproducibility of the combination across all supported configurations, with correspondingly broader post-market surveillance. Illustrative examples include dermal sensor patches and smartwatches (temperature, SpO2, heart rate) paired with a companion app.
Published: 2023-10-01
IMDRF
IMDRF/AIMD WG/N67 (Edition 1)
Machine Learning-enabled Medical Devices: Key Terms and Definitions
Edition 1 of the IMDRF Artificial Intelligence Medical Devices (AIMD) Working Group document on key terms and definitions for Machine Learning-enabled Medical Devices (MLMD). It establishes terms and definitions across the total product life cycle to promote consistency, support global harmonization, and provide a foundation for future guidelines on MLMD. Section 5 provides key definitions relevant to machine learning used in medical devices (including the definition of MLMD), Section 6 gives definitions from technical standards (e.g., bias, continuous learning, reference standard, reinforcement learning), and Section 7 discusses common machine learning terms. Most terms were previously defined in GHTF documents or internationally recognized AI standards, while some were developed or discussed by the AIMD Working Group. The document also notes that the term "bias" is used differently in data science and in legal discussions.
Published: 2022-05-09