The European Parliament has passed the first major AI act. The AI Act bans some AI systems or apps and puts restrictions and regulations on some. It depends on the risk associated with a particular AI system to decide which AI system gets more restrictions and regulations. The more risky AI apps will face more regulations and restrictions. 

Banned AI systems under the AI Act

Some AI systems that have very high risks associated with them are banned under this AI act. AI apps that include biometric categorization and facial recognition based on sensitive characteristics, social scoring, manipulating human behavior, and exploiting vulnerabilities are banned. 

High-Risk AI systems

Some AI systems or apps that fall into the high-risk category will have to follow strict regulations and face obligations like risk assessment, transparency, and human oversight. These include AI systems used in infrastructure, education, employment, medical devices, etc.

Obligations on GenAI and low-risk AI 

GenAI systems such as ChatGPT don’t fall into the high-risk category, but they need to adhere to some rules. They need to provide a detailed summary of the data on which their AI model is trained to comply with the copyright laws of Europe. Design the model in such a way that it doesn’t produce illegal content and reveals that the generated content is made with AI.

Spam filters and content recommendation systems fall into the low-risk category and have fewer regulations and restrictions.

When will the AI Act come into effect?

The act will take effect in three stages. It will become law by May or June.

After 6 months of lawmaking, countries need to ban the prohibited AI systems. After 1 year, rules for general-purpose AI will come into application. After two years, the act will be fully enforced.

If a company fails to comply with the Act, they can be fined up to 35 million euros, or 7% of their annual revenue.

Conclusion

The act will make sure that AI is only used ethically and that no wrongdoing with this powerful technology is done. At the same time, it is going to affect EU-based AI companies, and they will have to face issues while deploying their AI products. But it was necessary to regulate AI, and it sets an example for other countries around the world looking to regulate AI.

Read More

Apple Now Acquires DarwinAI to Strengthen On-Device AI

Copilot Pro Subscription Now Available Worldwide with a 1-Month Free Trial