Socially Responsible Game Education
In the last two months, there has been a marked upswing in public attacks against people, mostly women, who have either spoken out about sexism in games and the games industry or who want to make games that are a bit different from what is available to us now. My gut response to this phenomenon has been a combination of anger and helplessness. These attacks have even bubbled up to public discourse in recent weeks, which I fear has had the depressing effect of sending us back 30 years in the eyes of the general public.
Two weeks ago, I had the absolute privilege to speak on a panel about gender and games at Grace Hopper. Afterwards, I talked to students and heard some stories that only served to highlight a long-standing problem in games and higher education. I heard stories of students who are turned away from studying games because of its image problems, who are afraid of facing sexism in the industry if they can find a job, and who are afraid of speaking up because they don’t want to get shot right back down. These stories are not unique to one particular kind of game program, and I don’t believe any university is immune. I have heard similar tales in my own university and from students in other countries and everywhere in between. From students in dedicated game programs and students who might only take a one-off games class. This is the private face of gender discrimination in games. It makes me angry, too, but it doesn’t need to make me feel helpless.
Teaching Gender in a Non-Dedicated Course
I wrote a month or so ago about my plan to teach about gender issues in my Game AI class. My motivation at the time was to make sure that students were aware of the problems and had some vocabulary to discuss them, especially in the context of actually making games. I was deeply aware going into this that I don’t have many women in my class, and was worried about how it would go. I came out just as aware as before that I don’t have many women in my class, but also far more aware of how my students are thinking through these issues. Mostly, I came out convinced that this is an important thing for everyone to do, in every single games class. Women and members of underrepresented groups in games need to have a voice and feel like their voice can be taken seriously. It was obvious in my class that there was a lot of desire to talk through these issues and not much of an outlet to support discussion.
We speak a lot about how we want to bring more women into games and technology. About the beauty and excitement of our field, about how it will be strengthened by diverse voices. We make assumptions (that we could debate, but this post is already pretty long) that merely by having more diversity among those who make games, we will get greater diversity in the games that are made and sold. And yes, we need to fix up our pipeline, to extend it and to patch up the leaky bits. But we also have a responsibility to make sure that the people in our section of the pipeline are in a supportive environment. I think we put too much burden on the underrepresented to take on all these problems themselves (“lean in!”, we cry, as though the problem is that women don’t push hard enough, not that the system is always pushing back). The wonderful women who come out of our programs will go on to do great things and can be a force for great change in our industry if they want to be, but it is neither their sole responsibility, nor is the responsibility solely theirs.
I think everyone who teaches a game-related class has a moral obligation to address these issues with their students. Whether it’s in a town-hall meeting with all the students, or a 30 minute segment in your non-gender-related course, if students do not hear faculty speak out against gender discrimination and we do not provide a forum for students to talk through these problems in a safe environment, we are failing them. So here, I’m going to go over my “lesson plan” and the high points of our discussion in class on Games, Ethics, Politics, and Game AI. If I were doing this again, now, I would have perhaps had an exclusive focus on gender. We spoke about a mix of gender, violence, and censorship in class, which I think was still certainly productive and enlightening, but I think we could have spent an entire period talking just about how gender politics gets baked into AI systems and still barely scratched the surface.
With thanks to Jose Zagal for the suggestion, everyone was required to read the 1980 article “Do Artifacts Have Politics?”, by Langdon Winner. My course is cross-listed at the undergraduate and graduate level; graduate students were also required to read excerpts from Alison Adams’s “Artificial Knowing: Gender and the Thinking Machine” (a book I wish I had read much, much earlier in my studies). I chose these readings because I wanted the students to get a broader perspective on politics and technology outside of games, so that we could have a discussion without feeling like games are somehow special in the way that they encode political messages in their systems and skinning. The discussion ended up being quite good; with the barest of prompting questions around “why did you read this” and “how is it applicable to games” we got into a discussion of games and touched on gender, violence, censorship and localization issues, and software piracy.
Student consensus on the readings, though, was that they couldn’t see the immediate applicability to games, and also found the material very dry. My class is taught in computer science and as a result has students who may not be as accustomed to such readings as those from more of a humanities background. In class, the discussion ended up very good, but in the end-of-week reading responses, not as many picked up on the parallels between what they’d read and what we’d talked about.
There were three main themes that came up in our discussion of games, politics, and ethics: violence, sexism, and legal issues such as the boundary between localization and censorship, and software piracy. We touched on a lot of issues within these themes, in a jam-packed 100 minute discussion. There is room to turn many of these discussion points into an entire semester-long course! But I am glad that we aimed for breadth in the discussion, because it meant that we could get a lot of perspectives on a lot of different aspects surrounding these issues.
Violence. This was the first topic that came to mind from students. Why are games so violent? Is it a societal problem that games are violent? Can games be beneficially educational but not be teaching violence? What is it about the underlying systems of games that promotes violence? Getting into more concrete Game AI territory, we had some discussion about how a lot of the AI examples we’ll use in class are around violent video games – when NPCs should shoot with a shotgun vs. a rocket-launcher vs. a pistol, or maybe they should use a sword or punch instead. This is something I work to avoid in class, but it’s hard to avoid, and even when trying to avoid it I end up using those examples because they are the first that come to mind. So we also talked about why that is – that combat is easy to model and doesn’t require AI, and about other game mechanics that could arise from the ability to model intelligent behavior better, such as games about relationships.
Localization, Cultural Sensitivity, and Censorship. This was the surprise to me, something I have personally thought very little about but that my students brought up as an interesting place where politics and games collide. There are countries that have social norms that do not support some of the games we play in the US. Germany has strict rules about depictions of violence and especially of the Nazi party. China has strict rules about the depiction of undead characters in games. There was a lot of debate about the responsibility of game developers for localization, and where the line sits between cultural sensitivity and censorship. Not as much concrete discussion about how it fit into game AI here, but I think it raised overall awareness that when you make games, you are making them for people around the world, not just you and your friends, which is a great perspective for students to keep in mind regardless of the class.
Gender. We had a lot to talk about with regards to gender and games, which broadened out somewhat to all underrepresented groups in games. Some of the open questions we addressed were:
- Character creation tools. What is the default character in terms of gender, sexuality, and race? What options does the player have to alter their character – skin color, body type, sex, gender? This was a nice opportunity to discuss the difference between sex and gender. If a character creator doesn’t explicitly give players a choice of gender, are there indirect ways of expressing different gender identities? Does the game system respond to those choices? We spent a lot of time looking at Redshirt, an Indiecade finalist game whose character creator offers players a choice with a “slider” with male-ness and female-ness as extremes, as well as a sexuality option where bisexuality is the default. We used Redshirt’s character creator to talk about the different options you could give: whether it is an improvement over existing creation tools, what the “ideal” character creator might look like, and how that’s related to the game you are making it for. We spoke about how choices made in character creators are reflected in gameplay, with NPCs responding differently based on the characteristics you choose. A student asked why, if it is possible to make a character creator more flexible and open, one might choose not to: this itself opened a conversation about when and why flexibility in character creation is appropriate.
- “Normal” and “default”. Unsurprisingly, the conversation quickly turned to the dilemma of what to do with “female characters” and “female players”. How can you write AI for “female characters”? Let’s make sure that “female characters” aren’t always given gendered roles. How do we make sure that “female players” don’t feel alienated? In contrast, whenever referring to male characters or male players, there was no special descriptor. The conversation was always about “players” vs. “female players” or “characters” vs. “female characters”. This made for a good opportunity to ask questions about why it is that male-ness is the default, and what other “defaults” we may be unfairly considering.
- Scripted NPC behavior. Diversity among characters was a great way to talk about writing NPC dialog and scripted NPC behavior. Building on the conversation about defaults and norms, we talked through problems of authoring and how that relates to giving the player a choice in their avatar. For example, authored dialog that is unintentionally gendered because of assumptions about the player (the example I like to use is NPCs who compliment a player on their choice of “trousers” even if they are wearing a skirt), or authored behaviors that you might want to be different based on player/avatar identity. This let us talk a bit about how NPC behavior can be made modular or parameterized, which was a nice lead-in to the unit we were about to start on approaches for authoring NPC behavior, as well as whether and in what cases you want to have different NPC behavior based on gender.
- Games about social interaction. Finally, discussion of representing different personalities and identities in AI led us to being able to talk about one of my favorite topics: AI-based game design. We spoke about innovative games that are dependent upon social interaction (again, using Redshirt as an example). Given that we had this class period two weeks into the semester before we’d really covered any technical topics, this ended up being a really nice way to speak broadly about alternative uses of game AI, beyond controlling enemy NPCs that are trying to shoot the player.
Reflections on Teaching
Just as all artifacts inherently carry political values, all of us inherently “push agendas” when we teach. In all my teaching, I hold equality as a core value and want to encourage students to think critically about it. I attempt to ensure that women’s voices are heard through readings and through the games that I talk about in class. Most of all, whether it’s on the topic of feminism and gaming, or which approach to use in their AI systems, or both: I want students to learn how to explore questions that have no simple or clearly correct answers. I feel these values are best conveyed through discussion-oriented classes, not lectures.
Our discussion about gender became rather heated at times. There was a diversity of opinions in a not very diverse class, and many of the students lacked the vocabulary to really talk about some issues. The result was some insensitive statements that I don’t think were always intended as such. This can make it very hard to maintain an environment where everyone is on the same page and having a productive conversation. Some students responded with (often justifiable) anger to some of the more controversial statements made in class, and the discussion threatened to go well off the rails several times.
I proudly identify as a feminist. I’m still learning, though, and my identity does not mean I come automatically equipped with the language and ability to host complex discussions about it. My primary scholarly work is not in feminism and games (though I would like to branch out) but in procedural content generation and computational creativity. Computer science is not really a field where discussion-based classes are common (at least, in the many courses I’ve taken), so it’s taken me time to work out how to effectively lead classroom discussions. But here are some of the (probably obvious) ground rules I followed:
- I took effort to restate what students were saying in different language whenever it seemed like we were just going around in circles.
- When I needed to break discussion because things got heated, I attempted to make sure I was breaking it after different students each time, so nobody felt like they were constantly being silenced.
- I acknowledged that it is okay to have an emotional response to what was being said.
- I intervened only when needed to move points along, to tie together thoughts, or to point out when students were engaging in practices that others might find offensive or inappropriate.
It took a lot of effort to attempt to restate what students were saying in language that was clearer, to choose appropriate times to intervene while trying to make students feel like they were being heard, to draw a line between when it was fair to state an opinion and when what was said was unacceptable. There are a lot of complex, inter-related issues and I think we managed to touch on all of them, though sometimes in a roundabout and not terribly efficient way. All of my classes are discussion-oriented, and this experience was exhausting on a completely different level than other discussions I’ve led. But it was also an incredibly useful exercise not just for the students, but also for myself. I came out feeling more personally connected to my students, having learned how they are thinking through controversial and sometimes quite personal issues, and with a better sense of how to talk about these topics with people who don’t think about them all the time. I feel like some students are more open with me now than if I had simply held a lecture on finite state machines.
I have tried to make sure that in my class I am highlighting a diverse set of games that come from a diverse set of designers. I need to be constantly aware of this, and sometimes I have slipped. My syllabus certainly needs help; while it’s not terrible, I need to work to make sure I am including a diverse set of voices in assigned readings. I like to bring in guest speakers, but so far this semester, all of my guest speakers have been white men. I need to do better.
For this class period in particular, I would change a couple of things. First, I think Winner’s article is important and helped us frame discussions, but I would consider making it required for graduate students (optional for undergrads) and add in a game-specific article for all students to read. Second, I would have asked students to come into class with at least two examples how they think specific games have encoded gender in their AI systems. They came in from the readings completely unaware of what we’d be talking about in class, and we hung too much on the examples I’d really come prepared to talk about rather than looking to other games, and it made it feel at times more like a lecture and less like a discussion.
When I set this goal of talking about gender in my Game AI course, I really didn’t know what to expect. I would have considered the class a success even if all we did was to talk about gender and games in the abstract. I actually wondered before starting if we’d all just end up leaving early. I’d really hoped that we would be able to tie some of these issues back to game AI… and overall I think we were fairly successful in that. By placing it at the beginning of the semester, we were able to use this class period to close out our discussion of what Game AI even is, and we touched briefly on several issues (injecting personality into characters, modular NPC behavior design, alternative gameplay) that we then revisited (or will soon revisit) in later classes. I intend to keep teaching about gender issues in all of my game courses, regardless of content. The landscape is constantly evolving; but even if it wasn’t, these issues are important enough that they bear repetition from class to class. We cannot fix issues of gender representation in games and the game industry overnight. There is no simple or clearly correct answer. Decisions about how women are represented in games cut across a variety of aspects of game design and development: from initial concept to code execution, and everything in between. I believe we must equip our students to think about these issues from every conceivable angle. I believe that we must work especially hard to give a voice to the students who may be afraid to speak up. I’m looking forward to doing it again.