Practical guidance on how organisations can explain AI-assisted decisions to individuals, particularly when processing personal data. 

Examines the use of impact assessments to support accountability and cultivate trust in Artificial Intelligence (AI) systems. It highlights the growing deployment of AI and the need for ethical development and responsible use.

Focuses on algorithmic impact assessments (AIAs) as a practical tool for mitigating risks by reviewing AI systems' objectives, design, and purpose. It compares AIAs favourably to third-party auditing and conformity assessments, which face challenges such as a lack of standards. 

The Model AI Governance Framework for Generative AI from Singapore outlines a systematic approach to managing the risks and fostering innovation in generative AI. It proposes nine key dimensions for building a trusted AI ecosystem, such as accountability, data management, trusted development, incident reporting, and security. The framework calls for collaboration among policymakers, industry, researchers, and the public to address concerns like misinformation and copyright infringement, and also highlights the need for content provenance, safety and alignment research, and AI for public good. 

The Personal Information Protection Commission in South Korea has announced new policy directions for the safe use of personal information in the age of AI. These guidelines aim to minimise privacy risks while enabling the development of the AI ecosystem. Key initiatives include establishing principle-based regulations, creating a dedicated 'AI Privacy Team', and providing detailed personal information processing standards for each stage of AI development and service. The policy also promotes public-private cooperation to create sector-specific guidelines and strengthens international collaboration to shape global digital norms for AI. To foster innovation while safeguarding data, the commission plans to expand R&D in Privacy Enhancing Technologies (PET) and introduce an 'AI risk assessment model'.

The European Commission document offers guidance on interpreting and enforcing Article 5 of the AI Act, which prohibits specific artificial intelligence practices. It clarifies the scope, objectives, and interplay with other EU laws regarding prohibited AI, notably focusing on AI that manipulates, exploits vulnerabilities, or engages in social scoring. Real-time biometric identification (RBI) is given special attention, outlining exceptions for law enforcement under strict conditions, such as prior authorisation and adherence to fundamental rights.

The OECD Council's revised recommendation on Artificial Intelligence (AI) focuses on promoting trustworthy AI systems. It lays out principles for responsible AI stewardship, encouraging inclusive growth, respect for human rights, transparency, robustness, and accountability throughout the AI lifecycle. The document urges governments to invest in AI research and development, foster inclusive ecosystems, and create interoperable governance policies.

The Voluntary AI Safety Standard, released by the Australian Government, offers guidance to Australian organisations on the safe and responsible implementation of artificial intelligence. It promotes best practices through ten voluntary guardrails applicable across the AI supply chain, focusing initially on AI deployers. The standard aims to mitigate risks associated with AI while fostering innovation, aligning with international standards and Australia's AI Ethics Principles. 

New York City's regulations regarding the use of Automated Employment Decision Tools (AEDTs). The rules mandate that employers and employment agencies conduct bias audits of these tools to assess their impact on different demographic groups, ensuring fair hiring practices. The regulations specify how these audits must be performed, including the calculation of selection and scoring rates across sex, race/ethnicity, and intersectional categories

Colorado's Senate Bill 24-205 aims to protect consumers from algorithmic discrimination stemming from AI systems. The bill defines key terms like "algorithmic discrimination" and "high-risk artificial intelligence system," focusing on AI's impact on consequential decisions. It establishes duties for both AI developers and deployers, mandating reasonable care, documentation, and risk management policies. 

The Government of Canada's Directive on Automated Decision-Making outlines the framework for the responsible deployment of automated systems within federal institutions. It mandates that these systems are designed and used to reduce risks and improve efficiency while adhering to Canadian law. The directive requires algorithmic impact assessments, transparency in decision-making processes, and quality assurance measures. Crucially, it emphasizes the need for human intervention, access to recourse options, and public reporting on system effectiveness.