What is the problem?
There is no single role bearing responsibility for AI ethics concerns in all organisations using and developing AI-assisted systems. This could allow for unethical practices to develop and prosper.
Who should act?
European Commission, European Parliament, European Council/Council of Ministers (in forthcoming AI legislation).
Promote ethical awareness in organisations through creation of AI Ethics Officers to oversee AI ethics within organisations using and developing AI-assisted systems:
- AI developers and user organisations should have an internal position that combines scientific and technical understanding with ethical and human rights awareness to advise on the development and potential use of AI systems.
- Similar to the Data Protection Officer (DPO), this position should consist of a single, identified point of responsibility for overseeing compliance with and developing a culture of preserving AI ethical and human rights.
- The role must have necessary independence to speak out to prevent or mitigate risks of harm (akin to DPO and in-house counsel).
- The role could be required in high-risk cases, and only encouraged in others.
- The role may be aligned and combined with existing positions, e.g. DPO, Chief Ethics Officer, CSR Officer, or Business and Human Rights Officer.
- Qualification standards and education pathways for this position should be clarified.
In creating and/or implementing the role, key considerations are:
- Criteria for success:
- Definition of the role in law
- Presence of role-holders in organisations.
- Definitions, mandate, and parameters of the role need to be very clear to avoid creation of a powerless/”lame duck” position
- Clear ethical guidelines are needed to direct and support role-bearers.
SHERPA has produced an overview of ethical and human rights concerns arising in relation to AI.
SHERPA has produced 10 case studies demonstrating how ethical issues arising from AI impact on different sectors of the economy.
SHERPA has created sets of guidelines for operationalization of ethics by design for developers and for users of smart information, further developed by the SIENNA project. These can be used as a foundation from which to develop guidelines to support the role of AI Ethics Officer..