The pigeon and AI have similarities in their learning process.

The most important thing in the AI is the sensor that it uses. The similarity between the AI and the pigeon is that pigeon always finds its pigeonhole by following its senses. The thing is that AI might make many very complicated-looking things. But they are all made similar way as a pigeon finds its pigeonhole. 

The AI or robot can record its route using inertia or GPS, and then walk back by using the same route. That thing makes the "self-driving car" travel back its route when it loses the control signal. If the remote control signal is cut. The car simply drives its route back until it notices the signal again. 

Things like robot footballers are similar systems to regular robots. The sensor gets an image of the football, and then the robot can calculate its route to that ball. The robot can get more parameters like temperature, air pressure, and wind along with dampness and other things. The computer calculates the power of the kick and the point where that system must aim the kick. 

The humanoid robot that kicks the ball is claimed better than Messi says UCLA developed the footballer-robot. The robot can play football. And I don't know, does it use the outside calculation units? In that model, the robot can cooperate with calculation centers.

In a study conducted by the University of Iowa, researchers found that pigeons share similarities with artificial intelligence in their learning process. By subjecting pigeons to complex categorization tests, the birds were able to reach nearly 70% accuracy through repetitive, trial-and-error learning. This form of associative learning, where connections are made between objects or patterns, is also utilized by AI systems. Despite being considered a lower-level thinking technique, associative learning allows both pigeons and AI to excel at certain tasks, challenging the perception that it is rigid and unsophisticated.(ScitechDaily/Bird-Brained AI: Pigeons and Artificial Intelligence Share Surprising Learning Techniques)

"UCLA claims its humanoid robot footballer is ‘better than Messi’ (https://roboticsandautomationnews.com/UCLA claims its humanoid robot footballer is ‘better than Messi)

But things, like drone swarms can share their calculating capacity with that kind of robot. Maybe that kind of robot does not seem very impressive. UCLA created Footballer-robots to give data to other robotics projects. There are billions of things that can cause a fault in the AI. And the key element in AI-based robotics is this: AI gets all information that it needs from sensors. Then it must decide what to do when it gets a certain type of information. 

That decision is made by selecting the database which has the most similarities with the situation, that its sensors deliver. The system must select the database by following some variables. There must be some kind of identifier that the AI can use for selecting the database.

The database that has the most details of the situation, that it sees will be selected. That kind of system is a little bit slow. That's why there must be some kind of ranking system for the most used databases. 

And, of course, there must be databases that allow common reactions. In that system, the system might have a database that orders the robot to stop if it must compare multiple databases. 

https://roboticsandautomationnews.com/2023/03/14/ucla-claims-its-humanoid-robot-footballer-is-better-than-messi/65840/

https://scitechdaily.com/bird-brained-ai-pigeons-and-artificial-intelligence-share-surprising-learning-techniques/