I edit journal and conference papers written in English by Chinese graduate students. I do not think my job is at risk because a program could do it, nor that it will be any time soon.
Certainly there are things a program could catch, in particular errors with articles and other simple grammar problems. A recent paper had "an algorithm... designed by Korean scientist" and "[my technique] is faster than existed methods", for example. A program could catch and correct the existed/existing distinction, and notice the problem in the first one. However it could not know if that should be "by a Korean scientist" or "by Korean scientists" without either considerable contextual knowledge or a lookup on a citation (which one?) to see if there were multiple authors.
Until programs start routinely passing the Turing test, they will not be able to do the whole job.
Slashdot Mirror
I edit journal and conference papers written in English by Chinese graduate students. I do not think my job is at risk because a program could do it, nor that it will be any time soon. Certainly there are things a program could catch, in particular errors with articles and other simple grammar problems. A recent paper had "an algorithm ... designed by Korean scientist" and "[my technique] is faster than existed methods", for example. A program could catch and correct the existed/existing distinction, and notice the problem in the first one. However it could not know if that should be "by a Korean scientist" or "by Korean scientists" without either considerable contextual knowledge or a lookup on a citation (which one?) to see if there were multiple authors.
Until programs start routinely passing the Turing test, they will not be able to do the whole job.