Background
24F - Studio 1: Specialized Design Project-I&E Concept and Design
Partner
Valentina Pan
Duration
5 weeks
Introduction
Prompt
Using the LLM-Cabin framework, use your own product benchmarking framework to define and analyze a research subject concerning AI models in automobiles.
Background
Because my partner and I are both international students, we decided to tackle this prompt from a culture POV, to see how/if culture plays a part in users' preferences in AI models (specifically American, Chilean, and Chinese cultures.)
We brainstormed some prominent culture differences we see in our everyday lives such as: travel style, social etiquette, work-life balance, and the most obvious, a homogenous vs. heterogenous country.
Literature Review
“the people in interdependent cultural contexts (e.g., Chinese) are more likely to prefer AI as having capacities to influence, it follows that they are more likely to treat AI like they treat humans. This concept of "anthropomorphism" does not necessarily highlight the possibility that people in interdependent cultural contexts may treat elements of the environment (e.g.,AI products) as active, alive, and intelligent without likening them to “humans.” In other words, people in these cultural contexts may have schemas other than “humans” to perceive elements of the environment that can influence them” (Xiao Ge, Chunchen Xu, Daigo Misaki, Hazel Rose Markus, and Jeanne L Tsai. 2024.)
On the left is a proposed model of how cultures can be split into vertical and horizontal individualistic vs collectivist cultures. The model also splits the shown countries according to which category they would fall into.
We felt that Latin America as a whole was too big of an area to decide for Chilean culture, so we took a look deeper into Chile specifically.
Self-Construal Differences in Chile and South Korea: A Brief Report
Latin American and East Asian cultures are generally considered to be collectivistic cultures.This study intends to compare the self construal of Chileans and South Koreans and observe the variability in the presence of these dimensions in these two cultures. Chileans presented higher scores on independent self-construal than on interdependent self-construal, while Koreans did not show a preference for either dimension. These results are consistent with previous studies on Chileans, implying that not all Latin American countries would adhere to collectivism. (Benavides, Paloma & Hur, Taekyun, 2019)
Based on our findings, Chile did in fact not follow the Latin America trend as proposed above, but rather follows a Vertical Individualistic (VI) culture like the United States.
Hypothesis
Culture will affect people's preferences of AI models in automobiles. More specifically, US and Chile will prefer functions they can influence, emphasizing the feeling of control. China will prefer functions that will influence them, emphasizing the feeling of connection.
To build our own benchmarking framework, we looked into the following models: Horizontal vs Vertical Individualism and Collectivism scale and the Censydiam model.
Horizontal vs Vertical Individualism and Collectivism Scale
Censydiam Model
Regarding the literature on how culture affects choices, we found that many of them classified individuals and societies, so we chose this scale as one of the bases for cultural classification.
Because the previous model's classification of culture is more biased towards the whole society, and not detailed enough or linked to a user's personal needs, we found a model that also divides social and personal dimensions, which gives these 8 dimensions behind user motivation. It also investigates the impact of culture on user motivation and behavior, since different cultural backgrounds will lead to different needs and motivations.
We saw similarities between the two models since they both discuss the user's tendencies at the social and personal levels, so we decided to combine the two models together in hopes of seeing more clearly how culture affects user motivations. This model will be used as our benchmarking framework from now on. In our literature review, it was mentioned that collectivist culture wants the environment to influence people, while individualistic culture wants to influence the environment. Thus, the model is divided into two halves as shown below.
Framework
By matching the functions of a car's AI model to each quadrant, we came up with the framework below. According to our hypothesis, China would lean towards pleasure, connection, and belonging, while the United States and Chile would prefer functions such as control, recognition, and ability. To confirm our hypothesis, we used a questionnaire used in the first literature review on how culture affects people's views on AI, and matched the functions of a car to the eight dimensions of our model as cases for the respondents to choose from.
Interviews
We received a total of 89 responses from 30 countries, mainly from China, the United States and Chile.
The questions in the questionnaire are mainly divided into two categories. One is about establishing a connection with AI, which is marked in pink. The other is about controlling AI, which is marked in green. The bolded are the comprehensive responses from 30 countries. It can be seen that the results are more inclined towards control.
The importance of each aspect was also rated, and respondents believed that control was more important than connection.
The overall questionnaire results show that control is the most preferred function amongst all regions, followed by safety, recognition, belonging, pleasure, and vitality.
The map below is classified by region, and the graphs are drawn based on the 8 dimensions of our framework. Each color corresponds to a function.
Here we look at the degree of influence of each category. We mainly look at the proportion of blue (influencer) and purple (influenced). The pink belongs to the two functions which fall directly on the middle line of the two.
Findings
With China, we can see that the results are more biased towards "influenced", which is in line with our previous hypothesis. So when choosing a car, Chinese people will pay more attention to pleasure and a sense of belonging, such as the ambient lighting and immersive sound effects in the car, and functions that can help keep in touch with family and friends. They will also pay attention to some influencing functions, such as in-car control and the capabilities of the car model, but the proportion is relatively small.
In Chile, they tend to choose the functions of influence, which is in line with our hypothesis. However, there are also some functions that are leaning towards "influenced", such as pleasure and belonging. But overall, it can be seen that they are most concerned about control, including advanced safety functions such as driving assistance, and safety functions, like continuous monitoring of road conditions and warning of potential dangers.
With the United States, we can see that the results completely fall under "influencer". When Americans choose a car, they pay more attention to control and safety similar to Chile, as well as paying attention to the capabilities of the car model such as a seamless driving experience.
Conclusion
Looking at the three countries all together, we can see more clearly how they align with each other.
The United States (VI), is completely biased towards influencer, which shows that they do have a more significant individualistic tendency. Chile (VI) also has an obvious functional tendency towards influencer, however strays a bit towards influenced as well. China (VC) lies the most in influenced, preferring functions lying in that category.
When choosing a suitable automobile for these three cultures, we can infer that they would lean more towards cars with functions that highlight their preferred dimensions.
