The tech lobby wins in weakening the AI regulation

**European AI Act Amendments Signal Tech Lobby Influence, Delayed Implementation Raises Concerns** The European Union has reached an agreement on amendments to its Artificial Intelligence (AI) Act, a significant development that modifies the regulation shortly after its initial finalization. This adjustment, largely influenced by the tech lobby, allows for continued AI development with fewer immediate regulatory constraints. While this provides industry with more flexibility, it has raised concerns regarding the potential impact on public safety and fundamental rights. The core of the amendments involves significant delays in the implementation of crucial AI provisions. Rules governing high-risk AI systems, particularly those integrated into products, will now come into effect in August 2028, a substantial postponement from their original August 2026 deadline. Furthermore, regulations concerning autonomous high-risk AI systems are pushed back to December 2027. This extended timeline contrasts sharply with the rapid pace of AI innovation, leading to worries that essential safeguards for responsible and secure AI development are being postponed while the technology advances unchecked. Despite these delays, some protective measures remain intact. Providers are still mandated to register AI applications in a European database, irrespective of their perceived risk level. Additionally, stringent necessity standards will be maintained for processing special personal data, aiming to detect and rectify potential biases. A notable positive inclusion is the ban on "nudify" systems capable of generating explicit images of individuals, addressing a significant concern for privacy and protection against misuse, especially concerning women and children.