I hadn't really thought about the conductivity of the water.
I figured with the two electrodes close enough in a dry sponge. Any water would be enough to complete the circuit. I guess this needs to be tested lol. However, this design is so inexpensive, can't hurt to try. Any electrical suggestions are apprecited.
A few ideas off the top of my head:
The thicker the sponge, the more resistance when wet. So perhaps create a sandwich of aluminum window screen separated by a thin piece of fabric. This would allow creation of a large sensor.
Still, the wet resistance will be very high. You might be able to enhance it by pre-soaking the fabric in salt water, then allowing it to dry. If it gets wet again, salt water conducts electricity pretty well; though I'm not sure if this will reduce resistance enough to successfully deactivate the relay in the configuration you propose. As you said, it's a cheap experiment, so feel free to try and let us know if it works.
A possible further refinement: Build the screen/fabric sandwich, pretreated with salt, as described. Power one side with 12VDC. The other side goes to the base of a simple one-transistor amplifier. A low current will be enough to activate the transistor, which is connected to the relay coil. When activated, the relay will connect the hot and ground AC wires through a suitable resistor. This creates exactly the kind of fault that a GFI is designed to protect against. So if you replace your wall outlet with a GFI, and the leak detector triggers, the GFI will trip until manually reset; cutting off power to everything including the sensor. GFIs are recommended for aquariums anyway, as they protect you from other disasters.