“What if?”

The important constraint for me is that this project is about concentrated storytelling. The experience needs to start at the title, then as soon as they see the installation the story/world has already begun – their associations and assumptions about what is happening is already in play. For instance, with the game Journey. The use of an environment with the desert, sandy hills, flying, etc. These are playing with people’s biases (like the savannah preference) and the beauty of such movements.This is an important thing I learned when studying and playing with chatbots.

I think I actually discovered this insight from Brenda Laurel’s “Computers as Theatre” (which I notice is about to be released as a new edition). You can’t program a chatbot to handle every possible input a user may do in an appropriate way. So while you work hard to make the software as good as possible, work to capture as many keywords as you can and come up with clever response, what you also need to do is encourage the user to limit themselves. If you can’t answer everything they could possible enter, then you try and reduce what they could possibly enter. The way games do this is by giving you pre-selected inputs (choose one of these three or four options we’ve designed for you). But if you have a natural-language interface (where the user can enter in anything), then it helps to come up with ways in which the user with self-limit. If you created a salesperson chatbot, for instance, people would immediately limit their conversation to sales. But if the bot had no defined character or setting, people would input anything and therefore because the system couldn’t cover everything input it would stop making sense.

This is why I’m thinking of titles like “One of these robots will kill you”. I could set up the environment so it looks ominous – the robots are BIG and sharp edged, with quick actions, they have red eyes that follow you. There could even be a banner saying “Robot Kill School”. Perhaps the point then becomes – who is killing who?

The thing that is also in the back of mind as a concern is going with a title like “One of these robots will kill you” or “Robot kill school”. This works dramatically. And even though I’m creating a piece that aims to turn people assumptions around, I’m activating people’s negative bias. Their stupid biases. This is why advertising works, it targets our stupid biases. I don’t want to keep activating those biases, I want people to move beyond their biases. I want to talk to people who have moved beyond their biases. Negative biases are the lowest common denominator.

What if I went with ‘One of these robots will save you’? That is passive though. What about ‘You will save one of these robots’?… I wonder if it is better to have an action imperative that begins with what someone will do to you rather than what you will do to them? My instinct is that it is better to begin with what someone will do to you, because it triggers your instinct to act, to respond, to do what you need to do. Whereas starting with telling you what you will do triggers blocks: no I’m not! I won’t do that! You can’t tell me what to do! Etc. It stops the motivation to act before you’ve begun. Although if handled correctly, it could open up curiosity. Will this work really make me want to destroy a robot?


Leave a Reply

Your email address will not be published.