The UK Medicines and Healthcare Products Regulatory Agency (MHRA) updated its guidance on Software and Artificial Intelligence (AI) as a Medical Device (SaMD and AIaMD) on 3 February 2025. This revision includes a new link to guidance on Digital Mental Health Technology (DMHT): Qualification and Classification, clarifying regulatory expectations for AI-driven mental health solutions.
This update is part of the Software and AI as a Medical Device Change Programme, which aims to ensure clear, safe, and effective regulation of AI-based healthcare technologies.
Software as a Medical Device (SaMD) refers to software applications intended for medical purposes (e.g., diagnosis, monitoring, or treatment) without being part of physical hardware.
![](https://static.wixstatic.com/media/6516fb_58b95b021a264084a617292e944b081a~mv2.jpg/v1/fill/w_980,h_980,al_c,q_85,usm_0.66_1.00_0.01,enc_auto/6516fb_58b95b021a264084a617292e944b081a~mv2.jpg)
Artificial Intelligence as a Medical Device (AIaMD) refers to machine learning (ML) and AI-driven tools that adapt and evolve based on data patterns to support clinical decision-making or automate healthcare functions.
The latest February 2025 update introduces guidance for Digital Mental Health Technology (DMHT) to help manufacturers determine whether their digital mental health products qualify as SaMD under UK regulations.
This guidance covers:
Defining the intended purpose of a digital mental health app.
Qualification criteria for regulatory classification.
Risk-based categorisation to determine if the product falls under medical device regulation.
UK MHRA Software and AI as a Medical Device Change Programme
The SaMD & AIaMD Change Programme aims to create a more transparent and adaptive regulatory framework. Key focus areas include:
a) Roadmap for AI/Software Regulation
The MHRA outlines planned reforms across the entire product lifecycle, from qualification to post-market surveillance.
b) Pre-Market Requirements for AIaMD
Updated guidance on intended purpose and classification.
Introduction of risk-based AI assessment frameworks.
Focus on data transparency, bias mitigation, and continuous learning AI models.
c) Post-Market Surveillance and Monitoring
Ensuring real-world performance of AI-based medical devices.
Strengthening cybersecurity, bias detection, and system updates.
Intended Purpose Statements for SaMD
The MHRA now provides detailed guidance on defining an intended purpose for SaMD, ensuring regulatory clarity for manufacturers. Manufacturers must clearly define:
What the software does (e.g., diagnostics, monitoring).
Intended users (e.g., doctors, patients).
Usage environment (e.g., clinical settings, remote monitoring).
The MHRA outlines planned reforms across the entire product lifecycle, from qualification to post-market surveillance. For full details, visit the official MHRA guidance page:
Also, check out this standard about "Decision tree for navigating nanotechnology-based products for medical application" which is Scientific advice queries on nanomedicines have increased in recent years, particularly on lipid nanoparticles.
Comments