Countries around the world took steps to manage AI risks this week. Kenya released a detailed plan for ethical AI, including rules for data safety and fair algorithms. Their strategy encourages using AI sandboxes to test new ideas safely. In Europe, lawmakers pushed for stricter labels on AI-generated content like deepfake videos, while California passed laws punishing AI fraud.

The OECD introduced a voluntary reporting system for companies using advanced AI. Firms must answer questions about risk checks, data quality, and emergency plans. Over 50 companies pledged to join, with reports due by April 15. At the UN Global Conference, experts discussed AI in warfare, warning about autonomous drones and algorithmic hacking. They recommended global bans on AI landmines and stricter data controls.

Generative AI tools faced criticism for creating fake content. A study showed synthetic data from AI could spread false medical claims if not labeled properly. To fight this, researchers are developing detection tools and watermarking tech to spot AI-made images and text.

Major tech firms shared updates on AI safety practices. Google explained its AI lifecycle checks, including bias tests and hacker drills. AMD’s CEO emphasized ethics in chip design, ensuring AI hardware supports transparent systems.

Universities and labs received grants for AI safety research. Projects will explore model transparency and behavior prediction, aiming to prevent superintelligent AI from causing harm. Engineers were reminded to follow ethical codes when using AI tools, avoiding shortcuts that risk public safety.

Finally, bioethicists urged clearer guidelines for AI-generated data in science. They proposed honor codes and training programs to maintain trust in research. As AI keeps evolving, balancing innovation and responsibility remains a key challenge worldwide.