The EU AI Act is Almost Here! 

Via Unwind AI

“The EU AI Act is Almost Here!

After being proposed in 2021, hotly debated, and reported lobbying by big AI companies, the EU Parliament has finally approved the landmark AI Act with a majority. The Act employs a risk-based framework, categorizing AI systems into four levels of risk: unacceptable, high, limited, and minimal, where high-risk systems will need to adhere to the strictest requirements like risk mitigation, high-quality datasets, and human oversight in areas like employment, credit, and healthcare.

Key Highlights:

1. Companies are now required to notify people when interacting with AI systems, including chatbots and biometric categorization, and mandate the labeling of deepfakes.
2. The Act prohibits certain uses of AI, like real-time biometric identification in public spaces, emotion recognition systems in law enforcement and educational institutions, and indiscriminate scraping of biometric data.
3. A new European AI Office will be established to coordinate compliance, implementation, and enforcement of the Act.
4. The AI Act does not apply to AI systems developed exclusively for military and defense uses, focusing instead on civilian applications.
5. Companies face fines ranging from 1.5% to 7% of global sales turnover for noncompliance, depending on the offense’s severity and the company’s size.

“EU will pass world’s first comprehensive AI regulations

The AI industry’s Wild West era might be coming to an end. The European Union voted Wednesday to pass the world’s first major AI legislation. The bill regulates the rollout of new AI products across the union’s 27 member states, but it also provides a starting point for other governments who want to implement similar rules down the line.

Here’s what you need to know:

The bill passed by a significant margin: 523 votes to 46. The European Council is expected to sign off on the act in May. The first provisions will go into effect as soon as the end of this year. The bill asks overseers to rate AI products based on their anticipated threat level (a low rating would be a good thing in this case).

How does it work? Most AI-related tech, like spam filters, can expect to receive a low rating and will face relatively little oversight. But high-risk platforms — say, an AI-powered electrical grid — will have to seek approval before they’re allowed to enter the market. Other tech, including public face scanners and emotion recognition systems, have been deemed so dangerous that they’re banned altogether.

What else will change? OpenAI, Google, and other generative AI developers will have to hand over documents and data about how their models work. They’ll also need to abide by the EU’s copyright laws when training their LLMs.

Opinions on the legislation are divided. Some say the regulations will slow down progress and make it more difficult to achieve the holy grail of artificial general intelligence. Others argue the bill doesn’t go far enough in enforcing tougher safeguards that protect the public from misinformation and other threats.”