The European Union has published a proposal for a new set of laws regulating the use of artificial intelligence.

The proposal defines levels of risk for various applications of this technology. The highest level of risk ("unacceptable risk") includes social scoring by governments, exploitation of the vulnerabilities of children, use of subliminal techniques, and remote biometric identification. Under the proposal, these applications of AI would be banned.

The next level down ("high risk"), which would be heavily regulated, includes specific uses in education, recruitment, law enforcement and border control. Regulation here would cover ethics, data quality, documentation, transparency and accuracy. At this stage details are fairly sparse.

I welcome the proposals and for the most part they seem appropriate and much needed. However, I expect debate over this to rage for the next decade. The technology is rapidly developing and regulation will struggle to keep up.

🛎️ Why this matters: On regulation, where the EU leads, others tend to follow. It's early days: expect this to run and run.