It turns out, building robots is hard

even if you are only building them in a game.

Recently, I play-tested AI Escape (working title) in it’s most recent form, which was unfortunately incomplete. The issues that kept coming up with the game are as follows:

• Since the players were unfamiliar with the materials of the game, a master list of all the characters available should be included in the game. Whether as part of the game instructions, on a separate playing card like in Coup or a list for the players to share has yet to be determined.
• Chris’s suggestion that each character card be able to fulfill a robot build component means that I still need to determine what component each character corresponds to. I have since decided to scratch component cards completely and simply have character cards with the associated component listed below the design art under as a part of the character’s attributes.
• The problem of the mechanic that you cannot discard characters as this means an AI cannot be discarded. This needs to be stated upfront. Discards add an interesting element of variation to players’ hands however I have yet to determine how I can utilise it in my gameplay. I cannot keep track of how many decks are involved! (This is a problem in and of itself.)
• There needs to be an incentive for lying. One player took it upon himself to bluff for the sake of bluffing and to satisfy the game’s mechanics. Unfortunately, the players’ lack of understanding of the game meant that perfunctory bluffing actually inhibited gameplay.
• Perhaps as a product of the lack of incentive to lie, the game was actually not that enjoyable. I think this is because it has been bogged down in technical knowledge (jobs and components required to achieve the build as an example).

I have since decided to revisit the first proposed game idea – the conversation game in which each player is given a particular set of cards that prescribe what they can or can’t say based on human and/or AI attributes. The original post was entitled “We Are Dancers” and this will be the working title for the game. Brainstorming with Talen Lee we considered the idea of foregoing prescriptive words and using emojis instead. I like the ambiguity of this idea. I like how it links back to digitisation and cybercultures through the use of iconic symbols people encounter every day. I also like the possibility of players trying to interpret emojis from a machine’s perspective.

A variation on this idea is the inclusion of several AI characters as the specific attributes for conversation based on my research into various weak AIs (bots and personal assistants mostly) currently active or in production, for example, SIRI, Tay, Watson. I envisioned something like the Mad Hatter’s tea party from Alice in Wonderland. There would be an “Alice” player significantly outnumbered by AI players, all trying to interact with each other. It has occurred to me this could work like Cards Against Humanity with a stimulus card and individual characters’ responses (either prescribed on a single card and open to the player’s interpretation or given individual decks – the former option is much more affordable in production.)

My next post will focus on this new game idea, trying to incorporate whatever elements I can from the previous game idea and most importantly, production costs, budgeting, marketing and all that ugly reality stuff.