Not a robot at all
Given that each fly could act like a lighting implementation, researchers began experimenting with multiple pens, showing that multiple fly can be adjusted simultaneously. They also took a group of flies, overseeing each individually, switching between a smiley face pattern and a straight line formation.
Finally, the researchers found themselves placing the ball in the enclosure, leading the fly, and interacting with the ball and moving around. Once the fly began to interact with the ball, the researchers were able to direct the fly, sometimes moving it for more than a metre. There was nothing about the ball that was likely to be interesting for the fly.
In fact, there were many indications that flies were driven primarily by stereotypical responses to specific stimuli, but were not slaves to instincts. First of all, there was the fact that none of the experiments reached 100% accuracy. Furthermore, when multiple flies are guided through the experiment, they tend to roam the course more frequently when close to another, suggesting that they can alter their response to one stimulus when faced with competing stimuli.
All of this means that these animals are very similar to robots in most cases, but they are definitely not robots. There is a big difference between using electronics (and running AI models) and using the brain. Still, the researchers argued that remotely controlled flies have potential uses and determined that flies could carry milligrams of things, almost as much as their weight. Perhaps some simple sensors and electronics are enough. Obviously, the light control system is kept separate.
pnas2025. doi: 10.1073/PNAS.2426180122 (About DOI).
