Are you referring to the handwriting recognizer choices coming up with the simplified version or to the simplified character being converted to traditional in the Input Field?
If it's the former, that's actually just a quirk of the recognizer we use - not much we can do about it, sadly.
If it's the latter, are you using Input Field Compatibility Mode by any chance? If so, that mode forces characters to be converted into whatever the current encoding's character set is. If not, I have no idea what could be causing this...