The EU’s new AI law is poised to shake up the industry

A recent Stanford University study examining the ten most popular AI models in the world – including ChatGPT, Stable Diffusion, and Google-backed Anthropic – found each one would be in violation of a new EU draft law governing artificial intelligence.

🇪🇺🤖 Some quick background: The EU’s Artificial Intelligence Act, which still needs final approval from all 27 member-nations, would impose new requirements on any firm deploying AI systems within the EU – even if said company is located outside the bloc.

Among other things, the proposed legislation would:

• Require all AI-generated content to be clearly identified as such
• Impose mandatory safeguards to prevent AI models from generating illegal content
• Order AI companies to disclose any and all copyrighted material that was used to train their models
• Outlaw the use of AI in designated “high-risk” categories like facial recognition, predictive policing, or social credit systems

📝 So, what did the study find?... The Stanford researchers determined nearly all of the world’s top AI companies are in compliance with the EU’s proposed rule regarding the identification of AI-generated content.

However, nearly all of the top AI firms are not compliant with proposed mandatory safeguards against illegal content generation, or the law that would require them to disclose any copyrighted material used to train their models.

• The study’s authors concluded the EU’s new legislation, if fully enacted and enforced, would bring about “significant positive change” for the global AI industry.

📊 Flash poll: Do you think the US should adopt the same regulatory framework for artificial intelligence as the EU?

Yes

No

Unsure/other