Character-Level Tokenization

Spot the error when splitting a word into characters.

text = "hello"
chars = text.split("")

print(chars)

What is the bug in this character-level tokenizer?