Overview

• The EU's AI Act, effective since August 2024, reached its first compliance deadline on February 2, 2025, categorizing AI risks into four levels and banning 'unacceptable risk' systems.
• AI technologies deemed unsafe, including systems that manipulate, exploit vulnerabilities, or engage in social scoring, face severe penalties, such as fines up to $36 million or 7% of annual revenue.
• Proposed EU reforms aim to hold e-commerce platforms like Amazon, Shein, and Temu accountable for illegal or dangerous products sold online, shifting liability from consumers to platforms.
• The regulations reflect a stark contrast to the U.S., where the Trump administration has prioritized AI innovation with minimal regulatory oversight, including the rescission of a Biden-era AI safety executive order.
• The EU's regulatory efforts coincide with the rise of China's DeepSeek AI, intensifying global competition and prompting calls for ethical AI development and transparency from tech companies.