Critical Objects - Constrained Design Project - Facebook Trainer


As an exercise our professor placed a series of constraints on our design ideas for this particular project. Everyone in the class was required to design an object that offered a critic of some aspect of technology. Additionally, each project group was assigned a series of random parameters which their project would need to make use of. For myself and my partner Adi Dahiya our criteria were:
Attribute: Edible
Device: Humor
Mood: Satisfaction


Following a series of brainstorming sessions and discussions with our professor we narrowed the focus down to highlighting the dramatic monetary value our nearly every move on the internet generates for the myriad of data collecting companies. Focusing initially on the combination of the Edible and Humor criteria led us to think about developing some sort of device which would treat an internet user like an animal test subject in a primitive congnitive-behavioral study. The idea of conditioning an adult through the dispensing of desirable food the way one might influence the behavior of a rat felt ripe for exaggerated and absurd humor.

Continuing with this idea, we debated for a considerable time about whether a user should be “rewarded” for generating valuable data, or rather for avoiding this. While I do believe our concept could have worked following that latter logic, what we ultimately felt to be the strongest idea was to instead create an exaggerated fictional device concept which would condition users to generate even more data than they already do on a day-to-day basis.

It was at this point that our idea for the Facebook Trainer began to crystalize. Inspired by the quietly insidious new device from Facebook, the Portal, we created a concept for a fictional product which would hypothetically be a part of the same family of devices. This object, which we are calling the Facebook Trainer, offers you candy as a reward for specifically generating monetizable data for the Facebook company. We discussed a few different form factors for this device, including something like a desktop dispenser, but ultimately decided that a head mounted device would be most effective. By making the device hands free, it not only allowed the user to continue to browse the web while receiving dispensed candy, but it also essentially forces them to eat the candy via a tube running directly to their mouth.

While this concept felt strong, in a way it didn’t feel totally complete, and it was Adi who suggested that we may want to add some sort of negative feedback to actually train a user to stay off of websites through which Facebook is unable to track behavior. This lead me down a bit of a rabbit hole looking into the ways in which this sort of conditioning has been scientifically done in the past. What we essentially ended up developing was a device in the tradition of B.F. Skinner’s “Skinner Box” which would train animals to either complete a task to earn a reward, or complete a task to avoid punishment. I found multiple resources discussing the combined use of positive and negative feedback to reinforce behavior, and in particular the use of “mildly annoying but definitely not painful” auditory noises (quote from this paper by R.L. Herman and N.H. Azrin). Considering also we were already developing concepts that incorporated headphones, deciding to use noise as a punishment felt most effective as a component of this particular idea. We even went as far as to use a specific 3000hz tone as suggested for this sort of conditioning within the Handbook of Research Methods in Human Operant Behavior (page 106-107).

When it came time to begin fabricating, I started by looking at novelty beer hats for inspiration on mounting the candy dispenser and electronics to a piece of headgear. The dispenser itself went through various iterations inspired by different kinds of industrial candy machines, but ultimately it was the simplest design of a small paddle attached to a hobby servo which worked most efficiently for controlling the flow of the candy out of the receptacle. After doing further sketches to solidify the layout of the components, I began attaching each piece. While I had hoped to create a slightly more polished aesthetic, the constraints of the small size led me to simply using zip ties and glue, with the goal of a potential later version which would implement a more refined method. One component we were able to implement very simply was the inclusion of refillable/ replaceable candy containers. By mounting only to the cap of the candy container, we were able to make it easy to remove the receptacle and refill it.

The complexity of our project definitely existed more within the software side of the device. In order to accurately distinguish if a user was browsing a website containing facebook tracking or not, Adi developed a chrome extension which read each browser URL request and determined if it was directed to facebook or a facebook owned property (such as Instagram or WhatsApp). If you were on such a website, then the extension would send a message via serial to the connected MKR1010 microcontroller within the headset, activating the servo gate and releasing candy. if you continued to browse a website through which Facebook could gather your data, then every 30 seconds the user would be dispensed a new dose of candy. However, as soon as you visited a website without Facebook tracking, the 3000hz tone would play, getting louder over time to condition the user to switch back to an “approved” site and deactivate the tone.

in regards to the specific constraints on our project, the attribute “Edible” was obviously achieved through the use of food-based positive feedback. Watching the video of our device in action reveals the comedic absurdity of the concept which I believe fulfills the device “Humor” . And the act of performing a task (however inane) and then being rewarded for it, thus validating that action, undeniably serves as a source of satisfaction, thus including our final criteria as well.