Conscious Empathic AI

I and Hadi, have extended the initial Conscious AI framework to theorize the relationship between consciousness and empathy in AI and their impact on the service industry. The extended theoretical framework is published in the Journal of Service Research special issue of AI Service and Emotion.

In short, we argue that consciousness is neither an individual nor a social phenomenon but it is both individual and social. Consequently for consciousness to appear in AI agents we need at least two AI agents to co-create their own language. A kind of language that preferably has little to no roots in their training material and codes. For AI consciousness to be concluded then we need to observe AI agents repeatedly reaching the same agreement, through their co-created language, over and over to accomplish a task or series of tasks that they were not trained for. Additionally, utilizing extant philosophical understandings of the relationship between consciousness and empathy, we propose that conscious entities inevitability should develop empathic capabilities. Hence, conscious AI agents would be empathic and eventually develop moral responsibility at their full empathic performance.

Conscious Empathic AI is envisioned to revolutionize a range of industries relying on AI services and enable novel applications such as AI psychotherapy services. It is expected to lead to more adoption and use of AI in service as well as more AI accountability. We have discussed these possible applications in length within the article.

Picture References:

The digital art picture is taken from a France 24 article featuring our work in reference to Google’s LaMDA controversy.

The diagram summarizing our proposed theoretical framework is taken from the article publisher’s website