A video clip at Nature News shows a robot swarm marching into formation, creating a star shape. This is the latest achievement of the "Self-Organizing Systems Research Group" of Harvard’s Wyss Institute for Biologically Inspired Engineering. Their "Kilobot army" can create any predetermined 3-D shape on command. They call it "Programmable Self-Assembly." So far, the swarm has three shapes in its repertoire: a star, a wrench shape, and a letter K.

Each coin-sized Kilobot can sense the presence of its neighbor and respond accordingly. Notably, though, the video points out that "a user can give a desired shape to all robots" in the array (emphasis added). Sabine Hauert is wrong, therefore, to say in The Conversation that the robot swarm "assembles itself into shapes." Nothing would happen without the hidden hand of the "user" (i.e., the Harvard engineer). Is it appropriate for her to compare this to what nature does?

Self-assembly of this kind can be found in nature — from molecules forming regular crystals and cells forming tissues, to ants building rafts to float on water and birds flocking to avoid becoming prey. Complex forms emerge from local interactions among thousands, millions or even trillions of limited and unreliable individual elements. (Emphasis added.)

Hauert unwittingly leaps across the divide from non-life to life. She mentions crystals in the same sentence with cells, ants, and birds. Crystals are not programmed; they form because of electrical charges in constituent atoms that force them into some positions and forbid others. The living examples, by contrast, all have programmed information embedded in their DNA. The collective behavior might look like emergent self-organization, just like the Kilobot army’s march to the star formation might look magical to an onlooker. Nothing, however, was left to chance. The result was pre-programmed from the top down.

Other reporters made the same mistake, attributing self-organizing desires to the mindless robots. Elisabeth Palermo said at Live Science:

At present, the Kilobots are just working together to form 3D shapes — the letter "K" being their favorite. They can also transform themselves into common tools, such as wrenches and keys.

Like Hauert, Palermo compared the robots to animals that exhibit collective behavior. But don’t be surprised. The Harvard engineers themselves, publishing in Science, misled reporters by committing the same fallacy. They even attributed the bacterial flagellum and the Cambrian explosion to self-organizing principles in nature! Get a load of this:

In nature, groups of thousands, millions, or trillions of individual elements can self-assemble into a wide variety of forms, purely through local interactions. Examples can be found across a wide range of physical scales and systems: at the molecular scale with self-assembly of crystals or rotary motors of bacterial flagella, at the cellular scale with the development of multicellular organisms, and at the colony level with ants creating structures such as rafts, chains, and nests (bivouacs) using only their interconnected bodies as building material. Through collective shape formation, a group can dramatically change how it interacts with its environment. For example, the evolution of multicellular body plans enabled organisms to rapidly fill many ecological niches, and self-assembly of bridges and bivouacs allows army ant colonies to traverse difficult terrain while providing security and environmental regulation for the queen and brood. These examples of collective intelligence are fascinating to scientists across disciplines, as much of the global complexity arises from interactions among individuals that are myopic, sensing and interacting at scales many orders of magnitude smaller than the phenomenon itself.

To her credit, Palermo did notice that a lot of intelligent design went into the experiment:

This coordinated effort mimics the behavior of ants, bees and other insects that work together in huge numbers to build complex structures, such as colonies, bridges and rafts. But unlike bugs, these bots aren’t born team players. Researchers program each robot with advanced algorithms that enable it to move around on its own while simultaneously communicating with the other bots around it.

Her biomimetic description actually credits intelligent design not only for the robots, but for the ants, bees and other insects the robots are attempting to mimic. If the robots were programmed with advanced algorithms, who’s to say the insects were not?

Hauert shows how the robot experiment was not emergent, but programmed at a high level. The programmer’s goal permeates the whole system from start to finish:

First, all the robots are put together in an unformed blob and are given an image of the desired shape to be built. Four specially programmed seed robots are then added to the edge of the group, marking the position and orientation of the shape. These seed robots emit a message that propagates to each robot in the blob and allows them to know how "far" away from the seed they are and their relative coordinates. Robots on the edge of the blob then follow the edge until they reach the desired location in the shape that is growing in successive layers from the seed.

So far as we know, starlings in their murmurations do not communicate images of a "desired shape" to their neighbors, nor do swarms of anchovies, ants, or termites for their swarms, bridges, or mounds. This does not mean, however, that their behavior is any less programmed. The desired "shape" for living "individual elements" is function: thriving, surviving, building a habitat. If anything, the living examples imply even higher levels of programming, and more sophisticated specifications, to do what they do. The Kilobots are primitive by comparison.

One thing the living examples do share with the Kilobots is distributed code: the program is embedded within each individual, in its genome and epigenome. If a design inference is justified in the Kilobot case, it is justified in the animal cases.

Similar reasoning should apply to other claims of "self"-organization. Another paper in Science described "Self-assembled RNA nanostructures" produced by nanotechnologists in Denmark and at Caltech. Once again, the programmer is the hero of the story:

The aim of nanobiotechnology based on nucleic acids is to make three-dimensional objects with both precise architecture and function that can be programmed via their sequence. The ultimate goal is to build selfassembling, biologically active nanostructures within targeted cells by harnessing the cellular nucleic acid synthesis and assembly machineries. On page 799 of this issue, Geary et al. move us closer to this goal by designing RNA "tiles" that can fold into modular units as they are transcribed.

It’s all about having a goal, writing a program, and harnessing machinery. News from Aarhus University in Denmark emphasizes programmable code, not chance or self-organization:

"What is unique about the method is that the folding recipe is encoded into the molecule itself, through its sequence," explains Cody Geary, a postdoctoral scholar in the field of RNA structure and design at Aarhus University. "The sequence of the RNAs defines both the final shape and also the series of movements that rearrange the structures as they fold."

Indeed, these scientists had a hard time imitating the precise folding that living cells achieve so effortlessly. If it taxes their intelligence to sequence RNA to make it fold according to their wishes, why should anyone think that DNA works this trick by accident? The inference should be clear. If the world’s best human engineers struggle to imitate what life does, and succeed only by embedding their creations with complex coded algorithms, then it only makes sense that the superior performance seen in living examples came about by design.