Both of these
println 'test'.toBoolean() //false
println new Boolean('test') //false
instantiate a java.lang.Boolean using the constructor that takes a single String argument. According to the javadocs, the rule is:
Allocates a Boolean object representing the value true if the string argument is not null and is equal, ignoring case, to the string “true”. Otherwise, allocate a Boolean object representing the value false.
In both of the cases above, the String does not match ‘true’ (case-insensitively), so the Boolean created is false.
By contrast 'test' as Boolean follows the Groovy language rules for coercion to a boolean, which allows you to write:
if ('hello') {
println 'this string is truthy'
}
For a String, the rule is that if it’s empty or null, it evaluates to false, otherwise true.
I agree that this could be considered a bit inconsistent, but given a choice between consistency with the constuctor of java.lang.Boolean and utility, I think they were right to choose the latter.