The European Fee’s newest draft pointers present much-needed readability on how organizations ought to classify high-risk AI methods below Article 6 of the EU AI Act. Nevertheless, in addition they increase an essential query for enterprises: might your current AI methods already be thought-about high-risk with out you realizing it?
The reply could rely on greater than what the expertise does.
Underneath the EU AI Act, an AI system’s meant objective performs a central function in figuring out its danger classification. This implies how a system is documented, marketed, deployed, and used could be simply as essential as its technical capabilities.
Article 6 outlines two routes by which an AI system could also be categorised as high-risk. These embody AI used inside sure regulated merchandise and AI deployed in delicate use instances that would considerably have an effect on folks’s well being, security, or elementary rights.
For enterprise groups, this creates a number of speedy questions:
Which AI methods throughout the group fall inside the scope of Article 6?
Does present documentation precisely replicate how every system is getting used?
Might the Article 6(3) exemption apply, and what proof could be required?
What ought to authorized, governance, and expertise groups be doing now?
Airia’s on-demand webinar, EU AI Act: What It Really Requires and Enterprises Must Do Now, breaks down the brand new steerage and turns it right into a sensible choice framework.
The session covers the 2 pathways to high-risk classification, the constraints of the Article 6(3) self-assessment mechanism, and the steps enterprises can take to evaluate their AI methods extra confidently.
Entry the on-demand webinar to know what the most recent steerage means to your AI governance program and what your group ought to do subsequent.
