To discover how social Human-Agent Interaction is studied on the other side of the big pond, one of our group members - PhD candidate Rebecca Wald (Amsterdam School of Communication Research) - recently traveled to North America. Her first stop was Toronto (Canada), where she attended the annual conference of the International Communication Association (ICA). Right after, she continued her journey further south to State College (USA), where she visited Pennsylvania State University.
ICA23
What I most vividly remember from ICA - next to very sunny weather, great coffee chats, lovely street art, and yummy food - are three topics around social Human-Agent Interaction that came up in discussions across different conference events that I joined:
Concepts, concepts, concepts
What are we actually studying as Human-Machine-Interaction researchers and to what extent are different fields overlapping? Conceptual definitions build the core of scientific research. Before we are able to look at something very closely, we need to know what it is that we are looking at and what it is not. What a communication scientist means when referring to a robot might not always be the same as what a computer scientist means. This can be confusing and also quite exhausting.
What I took away from the discussions at ICA is that paying attention to conceptual detail is important, but we should not get lost in translation and let conceptual differences get in the way of our research too much. Concepts are not static and we must continue to reassess our definitions over and over again. I remember last year at ICA in Paris, I was very convinced that virtual assistants are quite different from robots. Now, after this year's ICA, I actually support arguments for why virtual assistants can be ascribed many robotic aspects. To me, this means we should always keep our research-mind open and accept that science is and stays messy (in a good way) :)
The award for the most popular actor in current research goes to... ChatGPT!
It was no surprise but still very noticeable: ChatGPT - the most recent AI language model developed by OpenAI - was present in all (!) conference sessions that I attended (and I tried my best to attend sessions across many different groups).
Are we all studying the same? What does this tell us? Is there an imbalance of scientific attention at the moment? Are we actually advancing the field if we all focus so strongly on one recent development, potentially neglecting others? These were questions I asked myself during the conference. I left ICA with the following answer: The strong focus on ChatGPT and alike is not a sign of imbalance or reason to worry. Frankly, I believe it is a very clear sign that there is something crucial happening in our society that is fascinating and scary at the same time and that touches upon many different areas of life. This is why ChatGPT receives so much attention. We, as researchers, follow our responsibility to investigate what possible pathways there are for this innovation to safely stay and grow with us. It means high responsibility - yes - but it also means empowerment. We are researching something very important and it gives us the chance to have a voice in ChatGPT's future (hopefully!).
Is 'being human' really the golden standard?
The third topic that I frequently encountered at ICA was about the standard that we regard as best when it comes to developing new systems that, in one form or another, interact with us. Should and can they be 'human enough'?
I heard reasons for both sides:
Yes, we need humanness in machines because research shows that this enables us to trust and engage more in conversations with them, which subsequently facilitates efficiency and success in multiple areas of life (e.g., customer support, education, medical care, etc.).
No, we should not generally see 'being human' as the golden standard for machines, as some systems can perform tasks that require aspects of human intelligence but they do not necessarily have to 'be human'.
Luckily, we have different camps present in the community, the 'YESs', the 'NOs', and of course also the 'IN BETWEENs'. We need those different perspectives to pause and question our own assumptions from time to time.
Visiting Penn State
After the obligatory trip to the very impressive scenery of the Niagara Falls right at the border between Canada and the US, I drove down to State College in Pennsylvania (USA). In this idyllic college town, I had the honour to visit one of the great experts in the field of human-AI interaction, Prof. Shyam Sundar, and his Media Effects Research Laboratory at the Donald P. Bellisario College of Communications.
My short visit fell into a quiet time on campus because of the summer period in the US-education system. This meant (almost) no students, calm streets, and a flexible agenda of the people I visited. And this turned out to be the biggest gift for me! Everyone was able and willing to take an incredible amount of time out of their days during my visit, allowing me to have the best week I could possibly have had.
Next to super interesting theoretical conversations about how humans interact with AI, joining lab meetings in person with people I so far had only met online, and presenting my dissertation-in-progress to get a fresh perspective on my thoughts, I was taken on a personal campus tour, on a nature hike, and out for amazing waffles as well as delicious dinner. At the end, I returned home with a notebook filled with thoughts, way too much new literature to read, and new contacts, but most importantly with memories I will hold on to for a very long time!
Not all sunshine and rainbows
In this blogpost, I tried to capture the most memorable discussion points from the conference and the greatest memories I made at PSU. Nevertheless, I also want to share that, in retrospect, going on a research visit right after a conference was not all sunshine and rainbows. It was tough at times. It drained my energy, tested my boundaries, and definitely forced me outside my comfort zone. My personal take-away is that after some big (work) travels, it might not always be possible to jump right back to normal routine. And that is okay. Creating space to navigate through challenges is just as much part of work-life as it is of life in general. In the end, there is gratitude for all of it - the good and the difficult.
留言