_This is a rewrite of the essay I submited with my college application in 2007._

A cloud of small, winged creatures darts across a calm green sky. They flash in a rainbow of gem-tone hues as the sun bounces off thier skin. Deep in the earth below, an ancient behemoth is sleeping. It’s abode is a damp, dimly-lit cavern; it’s bed is a mound of treasure and bones.

I admire dragons’ startling flexibility in stories. They are the ultimate actor, spaning the full range of roles. They have played both the protagonist and the antagonist, both the powerful and the weak. The physical definition of “dragon” is also loose: skin or scales or fur or feathers, normal vision or faceted bug eyes or no eyes at all, animal or human or alien intelligence, gnat-sized or cat-sized or house-sized or city-sized.

The enduring power of the dragon motif and the flexibility of thier definition seem to call for physical incarnation. There are already plenty of dragon statues; I want to go further. Rather than making a static idol, I want to make something that can actually be called a dragon. There are two main paths to acheiving this: construct a mechanical dragon or grow a biological one.

To go the biological route, you would start with DNA from existing animals, and combine it to create a novel one. The technology involved in this task would also be applicable to creating incarnations of other mythical creatures, such as chimera and manticores. While our collective understanding of how DNA and complex living systems work increases day by day, creating new animals is far beyond our grasp.

I know more about machines and electronics than biology, so I will more thoroughly consider the mechanical route. There are two major parts to this endevour: the body and the brain. In include in the body a standard amount of low-level intelligence: taking care of tasks such as moving in a given direction or staying aloft in flight. The brain portion of this equation refers to the higher-level AI that decides what to do, how to deal with the inputs received.

To begin with, let us more clearly define the problem of the body. The defining requirements here are flight and the ability to both land and take off. Generally, dragons have four legs and two wings; a long neck and tail add balance and a graceful profile. There are several modern approaches to flight – airplanes, helicopters – but they do not fit the image of a dragon. The wings will need to be those of an ornithopter, since they flap like a bird’s or a dragon’s. While we could try to make one of the greater behemoths of legend, targeting smaller-scale “dragonlings” would be both more useful and less expensive to develop. The fictional model for these dragonlings would be mainly Anne McCaffrey’s fire lizards.

It makes sense to start with the body and its intrinsic intelligence, since a specific dragon AI only makes sense if you have a dragon body for it to control. There are already remote-control ornithopters, so this task would be constructing a specific model that can carry both sensors and a micro-controller for both the body-AI and the brain-AI and writing the body-AI. There are already autonomous flying drones, although they do not use ornithopters to fly; writing the body-AI is a matter of creating a simple autonomous ornithopter-drone. Once there exists a body to respond to requests to “go that way” or “fly over to coordinates X,Y,Z”, it is time to consider the brain-AI. This brain is responsible to deciding what to do, and planning out the steps to achieve it; it would also be responsible for coordinating cooperative actions with other dragonlings.

This brain-AI would be best developed with a more specific task in mind. The most obvious first tasks would be various forms of surveilance. A group of dragonlings could sample the air quality over a larger area, to increase our understanding of how the chemical composition and pollen count change over time and altitude. They could also be used to quickly get a camera trained on some phenomena of interest; a lost hiker with a radio beacon, a leak in a pipeline, a reported fire.

If mechanical dragons were a reality, we would have a new tool for gathering information, for learning about our environment. They could also become a new part of our landscape – the comforting sight of shiny little dragonlings keeping an eye on things from above.