Why verification?
Verification is a tool to assess our AI system available to all to contribute to the use of ethical and compliant AI.
With this verification process, we discover or are certain about the risk category associated with our system in the pre-assessment phase, and subsequently, you know the details of the compliance with the seven core ethical principles of the PIO Model. Finally, a result is obtained through which we guide you and indicate the strengths and improvements of your system in accordance with the process carried out.
Why catalogues?
The catalogues serve as a tool for locating and learning about specific content included in the PIO Model. Both follow the order of the PIO Model, i.e. the categorisation based on the seven principles and the classification of the corresponding sections. The categorised information is also sorted alphabetically to facilitate navigation.
Text for the specific paragraph:
Legal: This catalogue contains the existing legal regulations and ethical recommendations that have shaped the questions raised in the model. In addition, each referenced regulation allows access to the corresponding text, to the specific article cited and to the PIO Model question where it is mentioned.
Examples: This catalogue collects examples related to the more than one hundred and thirty questions of the model, allowing you to filter your searches through real examples or concrete assumptions. In addition, you will find links to the journalistic or academic articles that supports or exemplifies the content, as well as the PIO Model question where it appears.
Why recommendations?
We provide ethical recommendation clauses for the purchase and procurement of AI systems which incorporate legal requirements and ethical standards and recommendations and serve as guidelines for public and private organisations to incorporate into their contracting and procurement processes.
These revolve around the seven basic principles of the PIO Model. They contain transparency clauses and require providers to provide full documentation of AI system design, algorithms and decision-making processes; fairness clauses that stipulate that AI systems must undergo bias detection and mitigation processes, demonstrating a commitment to fairness and equity in outcomes for all users; safety clauses that require rigorous risk assessment testing and implementation of safety measures to prevent harm; accountability clauses that insist on rigorous risk assessment testing and implementation of safety measures to prevent harm; accountability clauses that insist that AI systems must undergo rigorous risk assessment testing and implementation of safety measures to prevent harm; accountability clauses that insist on rigorous risk assessment testing and implementation of safety measures to prevent harm; security clauses that require evidence of rigorous risk assessments and implementation of security measures to prevent harm; accountability clauses that insist on clear vendor accountability frameworks, which should detail roles and responsibilities in the development and deployment of AI systems; privacy protection clauses that require compliance with data protection laws (e.g. GDPR) and implementation of data protection laws. GDPR) and implementation of privacy principles by design, safeguarding user privacy; autonomy preservation clauses to ensure that AI systems incorporate features that allow for monitoring, control and human intervention to maintain user autonomy; and sustainability clauses that mandate environmental impact assessment of AI systems and a commitment to social engagement practices that reflect the sustainability principle of the UN 2030 Agenda.
