Parsing the data with JSON is a safe and effective approach when the data is known to be well formed and parsable:
str = "[\"This is a word ect\", \"Char2\", \"This Element uses a (\\\") symbol that is important to keep\", \"Char4\"]"
require 'json'
JSON.parse(str).join(', ')
=> "This is a word ect, Char2, This Element uses a (\") symbol that is important to keep, Char4"
Otherwise a gsub solution using either regex or simple method chaining can be employed, but this sort of naive approach may remove quotation marks and brackets from the inner elements of the array string being parsed, potentially mangling the data you meant to extract.
str = "[\"This is a word ect\", \"Char2\", \"This Element uses a (\\\") symbol that is important to keep\", \"Char4\"]"
str.gsub('"', '').gsub('[','').gsub(']','')
=> "This is a word ect, Char2, This Element uses a (\\) symbol that is important to keep, Char4"
Notice how the gsub approach has a different result than the JSON parse method.
In theory, Ruby's eval could also be used to parse the string data into an array and then join it, but eval is meant for interpreting strings and running them as ruby code, and as such should only be used when it is important to run arbitrary ruby code that has been encoded as a string. The method name 'eval' actually comes from the word 'evaluate', not evil. Despite this, however, evaluation is not an objective in this scenario; parsing is.
Another reason why people are hesitant to recommend eval for trivial tasks like data parsing is that the worst case scenario of JSON#parse is that it fails to parse. Whereas the worst case scenario of eval is that you've completely deleted your file system by parsing a string that you didn't expect to be there when you first designed your code.