Z-Inspection View website

The Z-Inspection approach intends to guide innovators in the field of AI towards meaningful consideration and effective implementation of insights concerning ethical impacts of their developments and designs. It recommends a step-by-step process along the following points and provides tools and aids for each step:

• Pre-Conditions - Verify the pre-conditions, including the initial questions, the legal admissibility, and the absence of conflict of interests.
• Team - Form an initial team of multi-disciplinary experts.
• Protocol - Create a protocol (log) of the process to record information of the inspection process over time.
• Boundaries and Context - Define and agree upon the boundaries and context of the assessment.
• Analyze Socio-technical scenarios - Describe the aim of the AI system, the actors and their expectations and interactions, the processes, the technology, and the context.
• Identify Ethical Issues and Tensions - Build consensus among various stakeholders.
• Map to trustworthy AI - Map ethical issues and tensions onto the ethical categories established by the EU´s Guidelines for Trustworthy AI.
• Execute - Choose a strategy to perform the inspection. Define and execute Paths. Provide Feedback.
• Resolve - Address Ethical Tensions (when possible, resolve them). Give recommendations to relevant stakeholders. Start ethical maintenance if required.