Moral Decline With The Development Of Artificial Intelligence

When AI is starting to make decisions for us the question arises, what decision would a person make? Even more so, what drives us to make certain morally charged decisions and how are we affected by this? The first step in understanding is being able to differentiate a normal decision from a morally charged one, which is still a fairly new field of research. Up until now, the focus has been on things like ‘does one’s moral compass reflect on decisions in-game’, and, ‘how different personalities make different decisions’. And even though fMRI studies have been able to distinct morally charged decisions from normal decisions neurologically cite{greene2001}, more in-depth research into the physiological correlates and external tells (i.e. temperature fluctuations in the face) are lacking.

Moral decision-making is different from normal decision-making by the emotions that are involved; there is a sense of good and bad. This sense involves emotions which are found to be triggered in various regions of the frontal and parietal cortex cite{greene2004}. Also, moral emotions can mark emotional arousal over moral issues cite{turner}, which can influence moral decision-making. Being able to tell what emotions people are affected by at a given time is therefore important to be able to predict moral decision-making. To help add to the unknown, video games are becoming commonplace in moral decision-making research. For instance, citet{weaver} were able to predict what individual moral decisions people were going to make in Fallout 3. They did this by using Moral Foundations Theory by citet{haidt2012} which describes five pillars connected to morality. This theory will also be used in this study. Since we are only interested in the physiological tells when moral decision-making, Telltales The Walking Dead deemed more useful for us. According to citet{ryan}, this game is good for representing interpersonal, micro-moral scenarios for specific player-characters, which is what we are looking for. The game also has a strong connection to the real world and therefore players will feel more connected to the characters and scenarios.

To predict moral decision-making, infrared thermal imaging (ITI) was used, which allows for noninvasive temperature recordings. Using regions of interest (ROI), researchers have been able to observe certain emotions by recording cutaneous temperature changes in the face. Our goal here is to find which emotional profile is a predictor of moral decision-making. Finding this would eventually allow for machine learning algorithms to be able to distinguish moral from normal decision-making and allow for better access into human moral.

Some research has been done utilizing Moral Foundations Theory in video games research. For instance, in 2012 Joeckel et al. found significant correlations between pre-test Moral Foundations Questionnaire (MFQ) scores and in-game decisions using a game developed by university students. This was later supported by Krcmar and Cingel (2016) who had similar findings using the MFQ in Fallout 3. Weaver and Lewis (2012) were also successful in using Moral Foundations Theory in Fallout 3 and were able to predict in-game decisions as a result. Further, Grizzard et al. (2014) were able to influence scores on the MFQ by using a guilt-inducing game. After playing, participants scored higher on the moral foundations associated with guilt. Despite different games representing moral decisions in different ways, it is necessary to conduct research using other games. Video games lend itself best for this study since what we are after is an unconscious reaction which could be biased when using questionnaires or similar means. So far, no studies have examined the effects of moral decision-making in a video game context on cutaneous temperature changes. This study aims to fill the gap in the literature by investigating cutaneous temperature changes in the face when moral decision-making in Telltale’s The Walking Dead. Therefore, the following research question will be answered: “What effect, if any, does moral decision-making in video games have on cutaneous temperature changes?”

The goal of this study is to find the physiological/emotional correlates of morally charged decisions. This will be tested while people play Telltales The Walking Dead. Physiological responses, such as blood flow and changes in temperature will be measured with infrared thermal imaging. The experiment will test the following hypotheses:

No significant differences were found in the emotional/arousal profile in the face. There were two regions of interest (ROIs) which did show some potential but were not significant, these were the peri-orbital region and the maxillary area. Further, decision type (morally or normally charged), time, gender, and strategy usage had no interactions with cutaneous temperature readings. When disregarding decision type, time did have a significant effect on cutaneous temperature readings. This suggests that temperature changes in the face can be a predictor of decision making overall.