EU agrees to regulate potentially harmful effects of artificial intelligence

>The riskiest uses for AI are banned. According to the law, those include systems that exploit specific vulnerable groups, biometric identification systems for law enforcement purposes and artificial intelligence that deploys manipulative “subliminal techniques.”

So using AI for advertising, err I mean "subliminal techniques" is off the table now in EU?

I wonder what the definition of AI is?

Is all advertising at risk?