The MoralPLai project partners bring diverse perspectives: philosophical, technical and artistic, to shed light on the research topic, with each contributing valuable insights into the project’s key areas. Meet Jeffrey Schnapp, Carl A. Pescosolido Professor of Romance Languages and Literatures and of Comparative Literature at Harvard University, Faculty Director and Founder of the metaLAB (at) Harvard, Berlin & Basel and Faculty Co-Director of the Berkman Klein Center for Internet and Society,and read his reflections on the core research topics.
Why is it important to go beyond traditional research inquiry and dissemination approaches?
As a function of their canonical status, conventional modes of research and dissemination tend to operate within well-established limits and address well-bounded communities of expertise. Innovation that happens within such boundaries is invaluable, but tends to be less impactful and smaller in scope. Innovation that expands those boundaries is higher risk but also higher impact, engaging audiences that are unlikely to ever read a journal article or pick up a scientific or scholarly monograph, as well as creating a space for critical reflection on the future of discipline-bound modes of inquiry.
What has your fifteen-year experience as the founder of metaLAB (at) Harvard shown you about integrating the arts and humanities into scholarly, critical and creative practice to address complex issues such as AI literacy?
My experience first at the Stanford Humanities Lab (1999-2009) and now at metaLAB (at) Harvard (2011-present) has regularly reminded me of the rewards of bringing together deep modes of disciplinary and cross-disciplinary inquiry with thinking outside the box within a rigorous, project-based, hands-on framework. The word “laboratory” is important to me for precisely this reason: the laboratorium is a place of labor, of mental but also physical labor, a place where the mind meets the hand. Arts/humanities/science/tech integration has little meaning and is all too easy to embrace at an abstract level; it takes on a deeper meaning when tested within the practical challenges posed by concrete projects. If the projects in question are truly innovative they need to embrace an experimentalist ethos and embrace the risk of failure.
As AI technologies become more integrated into human decision-making, do you see a role for LLM-based chatbots in guiding people through moral dilemmas? If so, to what extent?
This is a complex question. Given the current state of the art, it’s hard to envisage a central role for LLM-chatbots in addressing moral dilemmas. Current LLMs are built on a “universalist” or encyclopedic knowledge base and, accordingly, pass some tests and fail many others. They regularly reveal the strengths and the weaknesses not just of their design, but also of the data sets they have been trained on. An LLM-chatbot trained on X and Facebook is clearly not going to be an authoritative source of moral guidance. Yet I can well envisage expert systems that are well bounded and whose data back end has been rigorously curated, designed to respond reliably to moral questions along the lines of various schools of reasoning or logic.
How can experimental and performative approaches, such as those used in the MoralPLai project, challenge conventional ethical frameworks and inspire new ways of thinking about AI’s role in society?
Experimental approaches in general and performative approaches in particular have the potential to test the actual social effects and impacts of AI in ways that are distinctive. They move us away from the realm of speculation, polemic and debate towards the realm of lived experience, allowing us to think critically and creatively about how AI feels as well as functions.
What is your vision and hope for the MoralPLai project after the performance in May?
I see this experiment as an excellent beginning. Time and multiple iterations of and variations on experiments of precisely this sort will help to refine a distinctive mode of inquiry which moves back and forth between the experimental to the experiential.
Visit the MoralPLai Project webpage to learn more.
