I used to meddle at robotics. There was an artificial neuron circuit many of us used to create simple "neural networks," Proffered up by Mark Tilden, they simulated the way a biological neuron worked from designing robots that sought or avoided particular stimuli.
One fella built a six-legged bug robot that had started acting strangely. It should have sought out the brightest spot in the room and not move until it "saw" a brighter spot. This way, if would follow a flashlight, or the sunbeam through a window as it made its way across the room
His, however, started acting erratically. I would sometimes move toward or away from him when he approached. It sometimes "patrolled" within a pool of light instead of sitting still. Other times, it would just sit there, scanning the room, and not move to the new bright spot until it was some distance away.
It was almost as if it had developed "behaviors" of it's own accord.
We troubleshot, we tested, we experimented. He couldn't reproduce it in a subsequent bot.
Then, somebody on our mailing list looked at the photo on the circuit board and realized what was up. A single unused input had not been tied to ground, and instead was now acting as a little antenna. the change of inductance from approaching it, or even from its own motors triggering, could trigger the neuron, causing actions that, from the outside, seemed almost like emergent behavior.
So, what the hell does this have to do with plants?
With apologies to Arthur C Clarke, a sufficiently complicated biological mechanism is indistinguishable from intelligence. When you boil it down, what is "intelligence" but the running state of a highly-organized glob of squishy on-off switches?