Owen Macindoe was one of a few IGDA scholars selected to attend GDC this past march. Here’s his report on the event.My time at GDC this year had a particularly serious tone for two reasons. The first was that I was attending as someone about to graduate and looking for a job in the near future. The second was that the industry was confronting an insidious culture of sexism that has grown in and around the games industry. It felt like both of us were doing some difficult growing up.
Being an IGDA scholar
I had been to GDC before as part of the MIT Game Lab staff, but whereas previously I had been attending to expand my research horizons and as a professional development exercise, this time around I was there for career reasons. I had been awarded the Eric Dybsand memorial scholarship through the International Game Developers Association and the AI Game Programmers Guild.
The scholarship was a tremendous boon for me. Not only did it allow me to attend the main conference this year, but through the efforts of the scholarship organizers, in particular Dave Mark, Steve Rabin, Luke Dicken, and Neil Kirby, I was introduced to a huge number of AI programmer peers and potential future employers. My networking schedule was non-stop and incredibly rewarding, thanks to them. I was very lucky to have been put in a position to meet all these people. For anyone considering applying for an IGDA scholarship, I strongly recommend it because it was a tremendously valuable experience both for myself and every other scholar I talked to.
The current state of game AI
On a professional level, the most interesting talks of the conference were the AI postmortems covering Hitman: Absolution, XCOM, Assassin’s Creed 3, and Warframe. Together these gave a good overview of where AI in the games industry currently stands.
The current state of the art in game AI is perhaps 5 to 10 years behind academia, which has at times been a bone of contention within both the academic and game AI communities, with developers seeing current academic research as impractical or irrelevant and game AI being seen as backwards and unsophisticated by researchers. However, AI is not unique in this respect. In general a 5 to 10 year lag for technology to reach industry from academia seems pretty standard.
Game AI for major titles has very hard limitations on computation time and memory usage, which makes directly applying many current AI techniques from academia, such as the decision theoretic planning work that I’ve been doing, impractical without a substantial amount of R&D work. Investing in that R&D is difficult for studios, since the financial stakes are high and research is by its nature risky. Some judge that the payoff for investing in AI research is not worth it.
At the same time, the incentives for research into game AI in the academic world are small. AI researchers in the US tend to receive funding from DARPA, IARPA, the NSF, or the NIH. Most of those institutions want to fund research that is going to make a real positive impact on people’s lives, such as improving health or education outcomes, or that have military or intelligence applications in the case of the defence agencies. Recreational gaming is not high on the national agenda, and tends to viewed as somewhat frivolous both by the funding bodies and also by the culture of academic AI. Serious and educational games have been a notable exception as of late.
In my opinion this is a great failing of the systems that support AI research. Since the 60s, games have been referred to as the drosophila of AI, and in studying them academic AI made great advances in search-based planning. Now in 2013 the problem of producing reasonable behavior from computers in real time and with very constrained resources is not unique to games. Self-driving cars, autonomous aircraft, and household robots are all actively being studies within academic AI and share some of the real-time and resource constrained properties of game AI. It seems to me like there should be overlap, and that new AI techniques could be proved out in games, and then applied in these other fields. The funding support for game AI research that might cross these boundaries is lacking, however.
Diversity in games
On a personal level, the talks that stood out above the rest this year were all in the advocacy track. The #1ReasonToBe panel with Leigh Alexander, Mattie Brice, Robin Hunicke, Kim McAuliffe, Brenda Romero, and Elizabeth Sampat was the stand-out session of the conference. It was distressing to hear the experiences of women within the industry and their reflections on the state of both the industry and gamer culture, but also hopeful to see people speaking out so thoughtfully about the issues at hand. This session, along with Anita Sarkeesian’s disturbing account of the online bullying that she received as a result of her Tropes vs Women Kickstarter project, and Anna Anthropy passionately reading her interpretation of Cara Ellison’s poem John Romero’s Wives, all highlighted the ongoing struggle within the games industry, and also the broader gamer community, to confront and the reform the undercurrent of misogyny that exist within it.
MIT Game Lab has been an active proponent of diversity within the games industry and the gamer community since the foundation of GAMBIT. Games from the lab, such as A Closed World, and community outreach through events such as the QUILTBAG Game Jam, have aimed at increasing diversity both through the inclusion of queer content in games and giving minority communities a creative voice within game development. In 2011, GAMBIT’s hate speech project documented some of the homophobic, racist, and sexist elements of the XBox Live community and online gaming forums that in 2012 have received a much more thorough critical examination through the gaming media. I personally am glad that, although 2012 was full of controversy over sexism in games and game development in particular, these diversity issues are receiving the attention that they deserve and there seems to be a cultural shift building that is slowly confronting and reforming the regressive elements of gamer culture. I am optimistic that we are on the cusp of a real positive change in the representation and treatment women and minorities in gamer culture and game development.