Consultation on the proposed “Organizational Framework for AI Accountability”

We are pleased to invite you to provide valuable feedback on our proposed organizational framework for AI Accountability.


As Artificial Intelligence advances, the need for clear and effective mechanisms to ensure accountability becomes paramount. Indeed, AI systems present unique challenges regarding accountability due to their complex algorithms, vast datasets, and self-learning capabilities. These attributes make it difficult to decipher the reasoning behind AI-generated outcomes, often leaving developers, deployers, and users in need of clarification about their responsibilities and obligations. In our work, we, therefore, want to support solving this problem by proposing a comprehensive and practical framework for organizations to clarify the duties and related measures for the risks that arise with AI systems to ultimately be able to define and assign appropriate accountabilities.

The project “Towards an Accountability Framework for AI Systems”

Our work is based on a collaborative research between the TUM IEAI and Fujitsu Limited. This 2-year project is dedicated to finding a practical solution to the accountability issues that arise with AI through research and synthesis of best practices and workshop-based exchange with industry and matter experts. Our findings resulted in the development of an organizational framework to detail accountabilities for AI-providing organizations.

In short, our developed framework proposes a solution to the definition of accountabilities by embedding AI ethics requirements at each step of the AI development lifecycle. Accountability is understood as taking responsibility and providing justification for one’s actions. Therefore, it is implemented in a risk-based manner, identifying risks for AI ethical principles at each step and ensuring prevention or mitigation measures. The core idea is to break ethical obligations down to more concrete actions for which responsibilities and, thus, accountabilities can be defined more clearly. This process is accompanied by ongoing measurement and monitoring to allow for faster reaction in case of potentially identified harm.

The Consultation Objective: Gathering Expert Feedback

We recognize that the success of our framework depends on its usability and practicability in real-world contexts. Therefore, we developed it in strong collaboration and regular consultation with a diverse group of industry and matter experts. We are pleased to now share the result of this work with a broader audience.

To further strengthen the framework’s impact in practice, we invite academia, policy, and industry experts to provide feedback on our proposed solutions. Your expertise and insights are invaluable in assessing our organizational accountability framework’s quality, usefulness, and effectiveness. We seek to ensure that our framework aligns with the needs of developers, deployers, and users across various sectors and contexts.

Therefore, we kindly invite you to join the consultation!

We encourage you to (1) review our framework summary and (2) submit your feedback through the survey by 8 Oct 2023 to share your thoughts, suggestions, and critiques. Your contributions will aid in developing practical and impactful solutions that foster ethical and trustworthy AI systems. We appreciate the time you will dedicate to this task.

Consultation Period

Start:  11 September 2023

End:  8 October 2023