In this 2021 publication, the authors introduce partial automation, an accessibility technique that delegates control of inaccessible game inputs to an AI partner. Partial automation works with other approaches to improve games’ accessibility, including universal design, player balancing, and interface adaptation.
The AI system offers players the ability to control the inputs that they want or are able to control (like actions, directions) and then provides input control for the parts they do not want or cannot control. A system like this can also offer balancing opportunities in matches of players with asymmetric skills (going beyond a handicap).
They tested their system with 6 players (some of them with severe motor impairments, such as tetraplegia). Participants reported feeling both enabled to play (they felt rewarded for their experience, which made gameplay feel meaningful and the automation let them focus on the most important aspects of playing the games) and disabled (mainly past experience with inaccessible games influencing their expectations and believing that gameplay was better for other people with different abilities).
Other feedback was that participants felt that the partial automation made it confusing to understand what was under their control and what was not.
Even predictable AI can be difficult to play with and the input ambiguity can make input frustrating occasionally. Partial automation seems like an interesting approach to support the accessibility of games though.
Read the full study
Gabriele Cimolino, Sussan Askari, and T.C. Nicholas Graham. 2021. The Role of Partial Automation in Increasing the Accessibility of Digital Games. Proc. ACM Hum.-Comput. Interact.
5, CHI PLAY, Article 266 (September 2021), 30 pages. DOI: 10.1145/3474693