Tefl.NET Forums

I apologize for the crude internet acronym but it's actually nicer than what I'm saying here IRL (in real life)! I just read this article from the Guardian about computers marking exam essays. When I visited the site for the Pearson Test of English Academic, I found that not only is the writing component marked by computer, but also the speaking! And they claim to be able to assess a person's fluency up to the CEF C2 level!

I am more than a little shocked by this, I really just don't believe that computers are capable of understanding and assessing non-native English as well as trained assessors, unless of course the essays and the speaking tasks are extremely narrow in focus.

I don't really have much to say, other than to ask how long it is until they introduce computer teachers and tell us that they are actually much more accurate and effective than human ones. :cry:

Here's a link to one of their FAQ pages. The test launches on the 26th of October.

Even using sophisticated applications that incorporate semantic search and compare technology, computers are still not at a level where language software of any kind is capable of entirely replacing a human element.

One possible advantage of the computer testing system is that marking would probably be more consistent, and less subjective, than humans when marking for style and content.

Personally I would not favour this Pearson system until I had had an opportunity to test it extensively for myself. I only use computers for checking MCQ type tests.