📋 Description
Adequate documentation is a key component for AI systems across their lifecycles and several risks arise when proper documentation is not maintained. Depending on the applicable legal and regulatory frameworks, a company’s AI systems may be subject to record keeping or documentation requirements. For instance, the EU AI Act requires certain entities to retain technical documentation about their high-risk AI systems. Additionally, robust documentation, such as an AI system's inputs, outputs, and decision-making processes, can address some explainability issues. Moreover, such documentation can help identify anomalies with outputs or assist with understanding an unexpected result.
A lack of proper documentation also raises issues around accountability. If an AI system makes an error or provides an output that results in negative consequences, then it becomes difficult to pinpoint the cause or assign responsibility. Ultimately, the lack of accountability can hinder efforts to fix the problem(s) or prevent similar issues from happening in the future.
