This ethics oriented service addresses experiments and applications that inherently have ethical risks. Particularly, when processed data is linked to human behaviour and status (even when anonymized) or AI learned models have direct or indirect impact a potential impact on individual or groups. Other cases can also be applications critical to the environment or infrastructures.
The concept of Trustworthy AI was introduced by the High-Level Expert Group on AI in the Ethics Guidelines for Trustworthy Artificial Intelligence and is based on seven key requirements:
1. Human Agency and Oversight;
2. Technical Robustness and Safety;
3. Privacy and Data Governance;
5. Diversity, Non-discrimination, and Fairness;
6. Environmental and Societal well-being; and
To assess those aspects and to identify potential risk for the experiment or the business cases resulting from the experiment, often very different perspectives are needed (technical, organisational, legal). This service, which can be provided by different or coordinatedDIHs (ITI, SDIL, Know Center), aims to bring those perspectives together in two on-site or online workshops and guides the SME through the process of identifying risk and possible mitigations.
SPECIAL ACCESS CONDITIONS
Particularly as the AI Act currently drafted by the European Commission sets out stronger requirements for providers of AI in many sectors, this service also helps experiments also to assess the regulatory impact on their business case. Adopting an “Ethics by Design” approach early might further increase the perceived product value independent of existing and upcoming regulations and generate trust. A Trustworthy AI Impact Assessment can be added as KPI for the first milestone as a result of the service (important pre-requisite is the involvement personell with technical, organisational and ideally data protection background at the SME during the workshops and afterwards)