https://www.cbc.ca/news/science/goldfish-driving-1.6309485
>>> Surprisingly, it doesn't take the fish a long time to learn how to drive the vehicle. They're confused at first. They don't know what's going on. But they're very quick to realize that there is a correlation between their movement and the movement of the machine that they're in.
As good as it sounds, I have my reservations about it showing "that a fish has the cognitive capability to navigate outside its natural environment".
Rather it seems to me the fish is just swimming in the direction of the treat, as it would do _inside_ its natural environment, and the car translates that directional push - I don't see a real "driving" coordination in the sense of a motoric abstraction based on a learning of cause and effect.
I doubt the fish are in any way "aware" of what's happening, or even that it's happening distinctly outside their environment. They're just swimming instinctively, adopting to what I imagine might a feel like a strong breaking current and a boop on the nose. Some might be more adept at this than others. After all they're captive animals that probably don't have much experience navigating challenging environments, not even aquatic ones.
I highly suspect "driving" via levers or buttons like the rats do would be beyond of what can be reasonably expected from a goldfish brain anyway.
As a kind of mental cross-check of the claim, I'd expect, say, a moth to be able to "drive" just the same way towards a light - and it would be pretty obvious there's no deeper cognition involved.