Components

The OCTAI project has four major components to it, which are part of a process of developing theologically-informed ethical contributions to the development and use of AI:

1. Listening to practitioners (in science, engineering, and industry)
2. Building a network of practitioners and theologians who are willing to collaborate
3. Producing theologically-informed ethical material
4. Proposing the Oxford Oath, a professional commitment for AI practitioners

A major focus across all four components is engaging with the majority world, building on the ability of the University of Oxford to bring people together from across the whole world, for the benefit of the whole world.

We intend to develop ethical reflections which attend closely to the ethical questions which arise in the daily work of AI practitioners. Scientists and engineers in the scholarly world and people working in industry are often faced with different questions in their practice from the conceptual questions which dominate public discussion. It is true that AI raises questions around human identity, which can be addressed through theological anthropology. The power demands of AI systems create significant issues around environmental impact. And the use of AI in weapons systems confronts us with questions around the proper use of violence.

But most ML engineers spend their days addressing much more granular challenges, and a key intent of this project is to understand the ethical issues faced by practitioners such as ML engineers. Offering theologically-informed ethics for practitioners requires us to first listen to practitioners, which we are doing through an initial series of workshops where practitioners (not theologians) present their daily work.

A second key intent is to build a network of practitioners and theologians who are willing to collaborate with each other. Because AI is rapidly changing the majority world, and yet the dominant voices in AI ethics are from the global West, we intend to use our platform at Oxford to surface the voices of the majority world to a wider audience.

We are building the network by invitation to those who are contributing to the work of OCTAI, with the intention of the network continuing to function and work in the future.

We intend to produce theologically-informed research outputs which are useful to AI practitioners. In the first instance, these will generally be traditional scholarly publications for peer review. We also intend to work to bridge from these publications to popular engagement, using the network to ensure that we achieve a wide distribution into the majority world and to practitioners.

Finally, we have already heard from AI practitioners that there is demand for a professional commitment, which we are suggesting could be called the 'Oxford Oath for AI.' While there are already regulatory and ethical frameworks, they function at a different level from an oath, such as the Hippocratic Oath in medicine. The type of commitment we intend to propose is analogous to relevant cognate initiatives that have been effective and influential, such as the Physicians Covenant and the Cambridge Declaration on Consciousness.  We hypothesise that taking seriously the values and vocational dimensions of AI research and leadership will facilitate the prioritisation of virtues in AI development and implementation. Producing a theologically-informed oath, even though we anticipate its form being non-religious, will allow us to attend to the need for ethics to engage with the growing majority of the world's population who are religious.