EU Considers Ban on Using AI for Mass Surveillance and Social Credit Scores

European Union flag
The law is essentially “GDPR for AI”

The European Union is mulling over legislation that would ban the use of artificial intelligence for several purposes, including social credit scores and mass surveillance. If the legislation passes, it will signal the EU taking a strong stance on how AI should be used – a stark difference from China and the United States. Some use cases would be allowed and policed similarly to digital privacy under GDPR regulations. 


For example, member states would have to create assessment boards to test and validate AI systems. Companies that develop prohibited AI in the EU, including companies based outside the bloc, would be fined 4% of their global revenue. 


A copy of the draft appears to include the following regulations: 


  1. A ban on using AI for “indiscriminate surveillance,” such as systems that track people physically or aggregate data from a variety of sources
  2. A ban on using AI systems to generate social credit scores that judge a person’s trustworthiness based on their predicted personality traits and social behavior
  3. Special authorization to use “remote biometric identification systems” such as using facial recognition cameras in public spaces
  4. Notifying people when they are interacting with an AI system unless it is “obvious from the circumstances and the context of use”
  5. New oversight on “high-risk” AI systems, such as systems that pose a safety threat such as self-driving cars or AI that can affect livelihoods, such as AI used in hiring and judiciary systems
  6. Assessing high-risk systems before they are put in place, including ensuring the systems are tested for bias and overseen by humans
  7. The creation of the “European Artificial Intelligence Board” with representatives from each member state. The Board decides which systems are “high-risk” and how to respond to them. 


Prohibiting AI use for surveillance and social credit is the standout section here. As good as the idea sounds, some privacy groups and digital rights activists feel it could be better defined – that the description of prohibited AI systems are too vague and open to interpretation. They argue it would be better for the law to be clearer.


The draft proposal may be changed before it is announced on the 21st. Even then, the EU may change it after getting feedback. It’s possible the articles will be clarified before becoming law – if it even becomes law. 

Leave a Reply

Your email address will not be published. Required fields are marked *

Previous Post
Surveillance camera with face recognition system

The DHS And ICE Both Use Clearview AI – Now They are Being Sued to Explain How 

Next Post
workers on strike in front of a factory. Factory closed, bankrupt

Spectrum Workers Build Their Own ISP While On Strike

Related Posts