Attack of the Robot Teachers: Artificial Intelligence and Automated Essay Grading

New Test for Computers – Grading Essays at College Level –


A computer program that can “grade” essays — of course, we knew it was coming eventually.  The idea makes me uncomfortable on a gut level, although it does seem like this program could have some interesting applications.  However, I am suspicious of the claim that a computer can truly “read” and understand written content on a level equal to — or as the creators seem to imply — superior to, a human reader.

According to the New York Times, the software supposedly uses assessment data provided first by live graders, which the program then applies to written content such as essays and short-answer responses.  The program will provide instantaneous feedback to students and allow for multiple revisions until students learn to “perfect” their writing.  This part of the program sounds promising as a formative teaching tool, especially if there are training modules that can be built in or that the teacher can provide alongside the program to prepare students for the exercise.  But comments on the program and its applications from its creators imply that the software can, and should, be used to replace human grading on final written assessments, an idea which I find deeply problematic.

My main objection to automated essay grading is that it pushes us further along the path to viewing writing as something formulaic, something governed by rules and patterns from which one must not deviate.  We have already blundered far enough down this path with our reliance on the standardized writing exam, which is part of most states’ compulsory annual testing, and which recently became part of the SAT.  Even with live, human scorers, a standardized writing test rewards students for adhering to a set formula, usually the dreaded five-paragraph persuasive essay.  This kind of testing is, and should be, the bane of any writing teacher’s existence because it saps the soul right out of what it is that we are teaching.  It makes students equate writing with drudgery.  It turns an art form into something perfunctory.

And yes, I suppose “good” writing can be created by following the rules and adhering to formulas.  But interesting writing breaks the rules well.  Brilliant writing creates new rules.  Meanwhile, a computer cannot comprehend originality.  It cannot compute irony, or even humor.  It can only look for and reward repetitions of the expected pattern.  And repeating a pattern, in my opinion, is not what teaching writing is about.  Teaching writing is about teaching the rules so students can learn to break them.  It’s about fostering creativity, allowing for experimentation, and providing a venue for human expression.  Unless artificial intelligence has truly reached the level of human consciousness, this is not something that even the most brilliant piece of software can provide.


4 thoughts on “Attack of the Robot Teachers: Artificial Intelligence and Automated Essay Grading

  1. Virtual_Momma

    I agree with you. Using a machine to help teach is alright, but not to grade assignments. It would take all the originality out of writing. Where would we be if everyone wrote, thought, and acted in exactly the same way? That is what it seems the government wants to do. What is the world coming to?


    1. esimmeth Post author

      I agree, VM — our current testing model encourages groupthink over individuality. Unfortunately, none of it is new — this has been in the works since the 60’s, according to the NY Times piece. My hope is that technology as it is currently used by some teachers and learners will help provide an alternative to the trend of standardization and uniformity.


  2. Jenny

    Oh, dear.
    And I thought the SAT writing portion was stupid (I’ve gotten an 8 on the essay every single time I’ve taken it. You would think that my essay-writing skills would have improved since seventh grade, but not according to College Board.).
    I love to run random school assignments through after I get them back. Probably not smart (who knows what thieves are lurking beyond the screen?), but it’s funny. It can’t even get the number of sentences in the assignment correct. At least they admit that their grading system is flawed. 😛


    1. esimmeth Post author

      Jenny, to me this is a perfect example of the problem with standardizing writing (what a concept!) — you are an amazingly talented, original writer and it is a complete shame that our current testing model promotes banality over originality. Write on and change the status quo!



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s