The US Department of Commerce has announced progress in AI safety, security, and trustworthiness, 270 days after President Biden's executive order on the Safe, Secure, and Trustworthy Development of AI. The National Institute of Standards and Technology has introduced multiple updates to support the objectives of the executive order, including the Generative AI Profile and Secure Software Development Practices for Generative AI and Dual Use Foundation Models. The department has also released an open-source software called Dioptra to evaluate the resilience of AI models against adversarial attacks. Furthermore, the plan for Global Engagement on AI Standards aims to foster international cooperation and the development of AI-related standards.

US Senators John Hickenlooper and Shelley Moore Capito have introduced the Validation and Evaluation for Trustworthy Artificial Intelligence (VET AI) Act to establish guidelines for third-party AI audits. The bill requires the Director of the National Institute of Standards and Technology (NIST) to develop voluntary guidelines for internal and external assurances of artificial intelligence systems. AI assurance is divided into two kinds: internal assurance and external assurance. The Director of NIST must develop voluntary guidelines for both kinds of AI assurance, addressing best practices, methodologies, procedures, and processes for assurance concerning consumer privacy, harm assessment and mitigation, dataset quality, documentation and communication, and governance and process controls.

Competition authorities from the UK, US, and EU have published a joint statement outlining potential risks to fair competition that can emerge from generative AI and the principles needed to support competition and innovation while protecting consumers. These principles include fair dealing, interoperability, and choice, with a focus on informing consumers about when and how AI is used in products and services. Agencies in the US and UK are cracking down on AI risks and becoming increasingly vocal about the need to ensure AI complies with existing laws and does not harm consumers.

The EU AI Act, which categorizes AI systems into various risk levels and outlines specific requirements and obligations, has been published in the Official Journal of the EU. The Act's implementation will be phased, with provisions concerning prohibited practices taking effect six months after the Act's entry into force. Compliance preparation is emphasized, and Holistic AI offers assistance in preparing for the AI Act.

Singapore released the Model AI Governance Framework for Generative AI in May 2024, which offers a comprehensive approach to managing the challenges of generative AI and encourages a global dialogue on the topic. The framework covers nine dimensions related to AI governance, including accountability, data, trusted development and deployment, incident reporting, testing and assurance, security, content provenance, safety and alignment R&D, and AI for public good. The framework stresses the importance of collaboration between policymakers, industry stakeholders, researchers, and like-minded jurisdictions. It calls for accountability in AI development and usage, responsible data governance, and democratization of AI access.