This white paper continues our series on trustworthiness in AI, examining various methods and frameworks for evaluating trustworthiness.
In recent years, there has been increasing interest in evaluating trustworthiness in AI, and several competing evaluation frameworks have been proposed. These frameworks differ significantly in regard to concrete evaluation methodology and level of technical detail. This white paper aims to provide more clarity on which framework can be suitable in a practical evaluation task. It looks at several published frameworks and reviews the dimensions of trustworthiness, such as fairness, reliability and safety, that they address.
