Automating Empathy, Globalising International Standards
by Project AEGIS/ Alexander Laffer, Lecturer in Media and Communication, University of Winchester
How do you make a truly global ethical standard? This is the RRI challenge that the AEGIS team have set themselves.
A bit of context: we are working with IEEE to create standard P7014.1, a piece of soft law that provides guidance and non-binding recommendations for those developing, using and regulating the use of emulated empathy in Human-AI partnerships.
This technology – often marketed as personal(ised) AI, assistants, companions, or even friends – is designed for sustained interactions between user and system. This raises myriad concerns, including the potential for manipulation of users and negative impacts on human relationships.
While these products are offered across geographical, cultural and regulatory boundaries, “global” governance tends to follow Western Liberal ideals and norms. However, we believe there is value in diversifying ethical frames of reference and are integrating diverse voices into our guidance to achieve this.
We start with our team: a mix of genders, ethnicities, and class backgrounds. However, we are largely UK-based, so we have partnered with international institutions, including the National Institute for Informatics in Tokyo, Monash University in Jakarta, and Data Lab, Tanzania, to broaden our understanding and engage experts and stakeholders from other regions.
We’ve conducted workshops with our partners leading to important insights for the standard. For example: differences in understanding of ethical and technical concepts; variation in views on how empathic-AI partners should function; as well contrasting levels of acceptance of AI-partners into social relationships.
The success of these workshops is supported by our commitment to Equality, Diversity and Inclusion. We have invited members of under-represented groups and worked to counteract structural hierarchies and legitimize the voice of all participants.
Our IEEE working group is also organized to ensure all members are given opportunities for meaningful input, for example, by adding meetings for otherwise marginalized time zones and forming sub-groups to address diversity.
By putting RRI (and EDI) principles at the centre of our work, we have taken a step towards improving the global ethical toolbox for AI. This is an ongoing process of learning international lessons from difference, understanding how cultures respond to technological propositions differently. No culture is innately right, but they contain useful knowledge and perspectives that are worthy of global adoption.
Please get in touch if you want to join us in developing a truly global ethical standard.
Visit Project AEGIS, follow them on LinkedIn or contact Alex Laffer.