Skip to main content

AI/ML Devices

Artificial intelligence (AI) and machine learning (ML) based medical devices represent one of the fastest-growing and most complex areas of medical device regulation globally. In Mexico, COFEPRIS is beginning to address AI/ML devices within its evolving SaMD framework, drawing on IMDRF and FDA guidance.

How AI/ML devices are currently regulated in Mexico​

There is no dedicated AI/ML device regulation in Mexico as of early 2026. AI/ML devices are regulated as software as a medical device (SaMD) where they meet the general medical device definition. COFEPRIS applies the standard three-class risk framework.

Key classification considerations for AI/ML devices:

  • What does the AI/ML software do? Diagnosis, treatment recommendation, monitoring, or administrative function?
  • How significant is its output in the clinical decision pathway?
  • What is the state of the patient if the AI/ML output is wrong?

High-risk AI/ML devices (e.g., autonomous diagnostic AI for serious conditions) are likely Class III. Lower-risk clinical decision support with physician override may be Class I or II.

Predetermined Change Control Plans (PCCPs)​

One of the key challenges with AI/ML devices is continuous learning — algorithms that evolve after market deployment. The FDA has developed a Predetermined Change Control Plan (PCCP) framework for this. COFEPRIS has not yet published equivalent guidance, but manufacturers with adaptive AI/ML devices should:

  • Document planned algorithm changes and their anticipated impact on safety and performance.
  • Assess whether post-deployment changes require a modification application to COFEPRIS.
  • Monitor COFEPRIS publications for emerging AI/ML guidance.

Clinical evidence for AI/ML devices​

AI/ML devices require clinical validation demonstrating the algorithm performs as intended for the specific intended use and patient population:

  • Algorithm training and validation datasets should be described.
  • Performance metrics (sensitivity, specificity, AUC) should be provided.
  • Subgroup performance (sex, age, ethnicity) should be addressed where relevant.
  • Real-world performance monitoring post-deployment is expected.

Equivalency route for AI/ML devices​

If an AI/ML device has FDA 510(k)/De Novo/PMA clearance or CE Mark under EU MDR, the equivalency route may apply, potentially with a 30-working-day review. The device must be identical to the approved version — which for adaptive AI/ML devices may create complexity if the algorithm has evolved since the foreign approval.