Lots of types of computer memory are well known and computers are able to use memory today to learn from examples.  Machine learning involves building up a representation in computer memory from lots of training examples.  The representation might be a neural network, a random decision forest, or another type of machine learning model.  Without a memory it would be very difficult to retain information from the training examples. 

Recently I read that scientists have found that plants are able to learn.  I was amazed by this because I had assumed that plants have no memory. 

Mimosa plants quickly fold up their leaves when they feel threatened.  Monica Gagliano, a scientist in Australia, slid mimosa plants in pots down a bar onto some foam.  The first two times the plants closed up their leaves but after two to six drops the plants started reopening their leaves as they learn it was safe.  

She put pea plants in Y shaped containers so the plant had to grow left or right.  A fan blew down one arm followed by light.  This was repeated for both arms and several days.  The plants learnt to follow the fan because this was a precursor to light the plant needs. 

I read that scientists think plants have memory which enables plants to learn.  But where is the memory in a plant?  Perhaps it is some type of distributed chemical mechanism which acts as a memory.

So what does this mean for artificial intelligence?  I think it means that there can be future types of learning machine which have plant type memories, which are distributed chemical memories yet to be invented in machine form.  The future of artificial intelligence is unknown and we will need to be careful not to define artificial intelligence in ways which will become inappropriate in future.