Jakob Ohlhorst (VU University Amsterdam), "Engineering virtue: constructionist virtue ethics"
Forthcoming, Inquiry: An Interdisciplinary Journal of Philosophy
I decided to try something new, and have ChatGPT interview me about my latest paper “Engineering virtue”. This was not as easy as it sounds because ChatGPT is quite obtuse. First, it just straight up produced a complete interview including my answers. Second, it gave me a catalogue of slightly boring questions. So on my third attempt, I had to explicitly instruct it to wait for my answer every time and ask its next question in response to the answer. This gave more or less satisfactory results. Still, I had to edit the output because ChatGPT summarised my response with exaggerated admiration, each time before asking the next question. Additionally, it began veering off-topic, and I doubt that it would have wrapped up the questioning on its own. For illustration, I left the second-last question un-edited.
ChatGPT: Can you give me a brief overview of your paper's main argument and focus?
Absolutely. In my paper, I argue that, in the past, virtue ethicists have just analysed what virtues were traditionally. Now, there is a new method – conceptual engineering – to instead develop new virtue concepts that better suit our needs. The great thing about new virtue concepts is, that if everybody starts using the new virtue concept, it will also begin to influence how we behave. In the long term just using the new virtue concept will contribute to making the new virtue into a real thing.
ChatGPT: Could you elaborate on the concept of "conceptual engineering" and how it applies to reshaping virtue concepts? How does this method work in practice?
Conceptual engineering is a new trend in philosophy. Instead of answering the question what our concept of something is, it aims to answer the question what our concept should be. This can involve both epistemic considerations – like which concept helps us best understand a phenomenon – but also ethical considerations – like which concept contributes most to everybody's well-being. Conceptual engineers usually take three steps: first, they assess whether a given concept fulfils its function well, second, they design a new or improved concept by proposing a new definition, and third they implement the concept by presenting it to people, encouraging them to use it, especially by explaining how the new concept improves upon its predecessor.
ChatGPT: Given your explanation of the steps involved, I'd like to explore the application of these ideas in the context of virtue ethics: How might the process of conceptual engineering be applied to the realm of virtue ethics, particularly in the context of developing new virtue concepts? Could you provide an example of how this process could be practically carried out within the realm of moral virtues?
In my paper, I use the example of courage. On a very traditional understanding, “courage” simply means military courage in battle. We, however, may want to honour also other behaviour as courageous which occurs outside of military contexts. We may for example want to consider it courageous for children to stand up to bullying. In that case, we'd have to change the archaic military definition of “courage” to also encompass much more harmless situations like a school yard. The new definition might then be something like “courage is a disposition to protect oneself and others even though one fears harm”. This obviously is a simplification, and standing up to bullying already counts as courageous.
ChatGPT: I'm curious about the implications of this approach for the broader field of virtue ethics: How do you foresee the integration of constructionist virtue ethics, with its emphasis on conceptual engineering, impacting the traditional landscape of virtue ethics? How might this paradigm shift influence how we understand and engage with virtues in ethical discussions and everyday life?
I must say that I think that this paradigm shift is already underway in the virtue literature. For instance, I take Miranda Fricker's proposed concept of epistemic justice to be a paradigm case of virtue engineering. What the case of epistemic justice shows, is that in contrast with traditional virtue ethics, we now have the conceptual framework to take into account formerly excluded perspectives. Additionally, our moral environment has changed, especially through the advent of social media; we now relate to others in novel ways. While our traditional virtues may sometimes prove useful, there are many contexts for which we need to develop novel virtues.
ChatGPT: Given the changing nature of our moral and social environment, could you elaborate on the types of contexts or situations where you think these newly engineered virtues might find significant relevance? Are there specific challenges or ethical dilemmas of our time that you believe would benefit from the application of constructionist virtue ethics?
Well, beyond social media and given that I am currently conversing with an AI, we can take this as an exemplary novel context. What moral and epistemic virtues do we require in interacting with AIs? What is an ethical approach towards artificial intelligences and how should we relate to them as sources of information? I can for instance not rely on traditional compassion towards an AI-model, but what would the appropriate version of compassion in relation to AIs be?
ChatGPT: The example of interacting with AI brings forth the question of how traditional virtues might need to be reshaped to suit emerging contexts. In this case, what adjustments or adaptations do you think are necessary to virtues like compassion, empathy, and critical thinking to ensure ethical and meaningful interactions with artificial intelligence? How might constructionist virtue ethics play a role in defining these adaptations?
I would argue that a key epistemic virtue for interacting with AI-models would be a novel kind of engrained understanding, AI-understanding, if you will. That is, the capacity to grasp what is going on in such a situation, to what cues an AI-model is sensitive, how information is processed etc. This AI-understanding must be different from our ordinary social virtues like empathy and understanding towards other humans or animals. Otherwise, the surface likeness between a large language model and human linguistic behaviour will be very misleading and make us easy targets for manipulation.
ChatGPT: The concept of "AI-understanding" seems essential for ethical interactions with artificial intelligence. How might this new virtue be developed and cultivated in individuals? Are there educational or experiential approaches that could contribute to fostering this AI-understanding? Additionally, how might virtue ethicists and society at large go about defining and promoting this new virtue in the context of AI?
The first step is to clearly define the goals of such a virtue, i.e. the ability to competently and reliably interact with AI-models. The second step will be to develop a conception of what is needed to reach these goals. Finally, we arguably will need some kind of training to acquire the relevant skills, like recognising AI-models, having a range of expectations about their outputs, and so on.
ChatGPT: As you mention, training plays a significant role in developing the AI-understanding virtue. How might this training be integrated into educational systems or even professional development programs? Additionally, given the potential for this virtue to impact not just individuals but also society at large, how might virtue ethicists collaborate with other stakeholders to promote the cultivation of AI-understanding on a broader scale?
Fortunately, there is already quite a lot of research on these issues. Virtue cultivation in both school and professional education is being investigated by a range of philosophers like Kristján Kristjánsson, Nancy Snow, and Jason Baehr. There even is the Intellectual Virtues Academy, a school that is focussing on encouraging intellectual virtues in pupils. There are also research projects looking into the understandability of AI-models. These strands will have to be brought together, so that we can develop and cultivate the AI-relevant virtues.
ChatGPT: It's heartening to know that there's already ongoing research and initiatives focused on virtue cultivation, particularly in educational contexts. The work of philosophers like Kristjansson, Snow, and Baehr, as well as the Intellectual Virtues Academy, highlights a growing interest in fostering virtues among students and professionals.
Given this collaborative approach, I'd like to explore the potential role of society and its engagement with this new virtue:
Society's understanding and ethical approach to AI interactions could significantly benefit from the cultivation of AI-relevant virtues. How might a collaborative effort between virtue ethicists, educators, researchers, and even policymakers help bridge the gap between the theoretical development of virtues and their practical implementation in society's approach to AI? What challenges or considerations might arise in this collaborative endeavor?
Isn't this question getting a bit far afield from engineering virtue? Shall we wrap up? Do you have any last questions?
ChatGPT: Certainly, my apologies for diverging from the core topic. Let's bring our discussion back to the concept of engineering virtue. Before we wrap up, one final question:
As we conclude our discussion on the concept of constructionist virtue engineering, is there a specific message or key takeaway that you hope readers will glean from your paper? How would you summarize the significance of this paradigm shift in virtue ethics for both academia and society at large?
For me, the main take-away of the paper is that we are not tied to the virtue concepts that we have been attached to since forever. We can, do, and sometimes should, think about how our virtues could be better. If we succeed at anchoring improved virtue concepts in society, then there will be considerable benefits downstream because we will strive to be better people.
ChatGPT: Thank you for your time and insights throughout this interview. Your exploration of constructionist virtue ethics and its applications has been enlightening and thought-provoking.