Embedded Files
The use of language enhancement applications (LEAs)
The use of language enhancement applications (LEAs)
The University has recently established a policy with regard to the use of so-called language enhancement applications in assessed student writing. LEAs are products such as Grammarly which work as browser plugins or standalone resources to offer suggestions on improved grammar use and word choice, and can include 'auto-fix' options. When attempting to assess a student's language proficiency level via written coursework, these tools can have a distorting effect, and it can be difficult and sometimes impossible to tell what is students' own work and what has come from the app.
In short, the newly-drafted policy permits the use of these applications in student writing, but makes a distinction between adopting language suggestions from a 'checker' such as Grammarly, and the use of more comprehensive tools such as machine translation software. The latter would be deemed academic misconduct as it would not by the student's own work. This policy has been incorporated into the policy as outlined in the course handbook.
Impact on teaching and assessment
Impact on teaching and assessment
The first step in adapting the teaching of the Language and Study Skills module to the world of LEAs is raising awareness with the students. One of my departmental colleagues, Eddie Cowling, produced a useful learning object which I have included in the scheme of work, and make reference to when teaching independent study tools, referencing and academic misconduct. The point is to enable students to use such tools judiciously, recognise their limitations, and help them remain in compliance with the department policy on acceptable use.
Technology has profoundly changed the way we write, which has massive implications for language teaching and assessment.
Another way that my teaching practice has adapted to this policy is in marking assessed coursework. Criteria used for assessing writing proficiency refer to students' range and accuracy of vocabulary and grammatical structures, and these are clearly affected by the use of LEAs. If we accept and expect a certain degree of LEA input in coursework writing, it is probably unrealistic to attempt to make fine grained distinctions between proficiency levels as I had been doing thus far. I created the simplified criteria referred to above to streamline the marking of pass/fail assignments; however, this simplification has also helped deal with the issue of LEAs in marking. The three bands available (fail, pass and strong pass) distinguish between intelligibility (whether it's possible to understand what's being communicated) in such a way that mitigates the impact of tools such as Grammarly. While these apps can fix errors and make the writing sound smoother, they do not commonly conjure complex structures and nuanced lexis without making the text sound slightly odd and losing the intended meaning. The original criteria which link to Common European Framework of Reference levels are still used for marking exams.
Reflection
Reflection
These changes to my teaching in response to the department policy on use of LEAs represents the the tip of the iceberg, and further changes in policy, teaching and assessment practice are inevitable as AI technology advances. The policy currently advises 'judicious' use and avoiding 'overreliance' on LEAs; however, these are fuzzy distinctions, and in reality it is difficult and sometimes impossible to tell where the 'suggestions' end and students' own work begins. While the use of LEAs in writing has implications for all departments when dealing with issues of attribution, language teachers are at the sharp end of this conundrum. What is the best way to accurately assess someone's language proficiency when we are surrounded by tools that mean we rarely have to rely on our own knowledge of a language in order to communicate?
Grammarly works by offering language suggestions to writers, and can be installed as a browser plugin.
Conferences such as BALEAP have been paying increasing attention to LEAs, and offer a useful opportunity for language teachers in higher education to discuss policy responses to the swift changes in technology.
For future development of my module and my teaching, I plan to make these tools more central to the curriculum. Groves and Mundt (2019) advise teaching students how to use LEAs and even translation tools to improve their own writing by comparing the suggestions they make to their own writing, and by teaching them how to evaluate the validity of suggested grammar fixes and translation. I am currently organising a day long conference on the topic of ownership, integrity, argumentation and authorial voice in language teaching in higher education. Mike Groves and Klaus Mundt, who have done extensive research on the use of translation software and LEAs in student writing will be running a plenary workshop that will look at ways in which we as teachers can adopt rather than resist the technology.
I believe higher education and language teaching especially have been slow to react to these technologies, and we need to formulate more robust and comprehensive policies as well as standardize teaching and assessment practice; however, I also think that adapting our teaching to the technology is going to be crucial in order for language instruction to remain relevant. I do not know what all the answers are at the moment (or even the right questions), but I am hopeful that by bringing this topic into the discussion spotlight, as I'm attempting to do with the upcoming conference, the teaching community will be able to formulate constructive responses that enable students to write proficiently in a way that retains their own voice, and to control the technology rather than the other way around.
Copyright legislation
Working with others
Page updated
Report abuse