A systemic study by the University of Cambridge urges tighter regulation of generative AI-enabled toys that interact with children and suggests introducing new safety kitemarks. The authors recommend that only developers adhering to specific guidelines should access generative AI models for toys, emphasizing the importance of recognizing the social and emotional aspects of early childhood development.
The Council has agreed on a proposal to streamline rules regarding artificial intelligence (AI) as part of the EU’s “Omnibus VII” legislative package. This initiative aims to enhance legal certainty, support innovation, and ensure harmonised implementation across member states while addressing key ethical concerns related to AI practices.
The National Artificial Intelligence Ethics Framework has been issued to guide the safe and responsible development and deployment of AI systems, ensuring they are beneficial to individuals and society. It mandates safety, human oversight, and the prevention of harm, while also emphasizing fairness, transparency, and the need to mitigate biases in AI operations.
In response to the complexities of AI-driven decision-making, Shekhar Natarajan introduces Trust Ecology, a new paradigm that embeds ethics and accountability directly into AI systems. This innovative framework aims to create trustworthy decision ecosystems by focusing on systemic integrity rather than individual attribution.
AI has created a 92% visibility gap in identity governance, undermining traditional control models as non-human identities proliferate within core systems. To address this, federated governance offers a unified control layer that encompasses both human and machine identities, ensuring accountability and policy-driven oversight across platforms.
Members of the European Parliament have reached a preliminary agreement on amendments to the EU Artificial Intelligence Act, which will be reviewed before a scheduled vote in Brussels. The proposed changes include extending compliance deadlines for high-risk AI systems and banning non-consensual explicit deepfakes to enhance consumer protection and online safety for children.
The draft procedural rules for enforcing the AI Act would allow the European Commission to access and inspect general-purpose AI models, including their source code and hosting infrastructure. Additionally, the plans outline independence criteria for experts and provide a 14-day response period for providers regarding preliminary findings.
Robotics companies are rapidly scaling AI capabilities, but many boards struggle to keep pace with adequate oversight. Essential questions regarding model risk, data provenance, safety cases, and incident response plans are crucial for directors to protect enterprise value while fostering innovation.
The European Commission has released a draft detailing the procedural rules that will govern investigations into AI developers. This draft is open for feedback until April 9 and outlines conditions for proceedings, file access, confidentiality, and penalties.
AI assistants struggle with governance, leading to inconsistent rule adherence. By implementing a structured governance layer with hooks and skills, Claude Code can dramatically improve task compliance and reduce errors in development.