On 29 September 2022, the public consultation closes on the second draft of the AI Risk Management Framework published by the National Institute of Standards and Technology (NIST). In its second draft, the NIST listed seven characteristics of "trustworthy AI" and the risk that the developers must address to meet the criteria in assessing the system's trustworthiness. The characteristics are "valid and reliable", "safe", "fair – and bias is managed", "secure and resilient", "transparent and accountable", "explainable and interpretable", "privacy-enhanced". The AI framework aims to improve the ability of developers to incorporate trustworthiness considerations into the design, development, use, and assessment of AI products, services, and systems through voluntary adoption of the listed principles and criteria.
Original source