Assessing communication skills in the age of AI

In the age of AI, communication is changing, but effective human communication skills are more important than ever.

Many educators would agree that university graduates will need to know how to use AI-mediated communication in their future professional contexts. How to prompt and edit AI to produce a precise and coherent report. How to create an AI-scripted presentation. How to refine the language of business emails. How to produce AI images and videos for architectural, design and engineering projects. And much more.

But AI cannot replace the human interpersonal communication skills required in most professions. The skills to critically edit and revise AI-produced text. To draw high-level context-specific conclusions based on AI-collated evidence. To skilfully pitch ideas and negotiate client needs. To comfort and advise a suffering patient or struggling student. And to effectively interact with other humans in team meetings and projects.

In a recent survey, communication ranked as one of the top ‘soft skills’ (also termed ‘power skills’) needed in the future workplace: interpersonal connections are more important than ever.

So how will universities ensure that students are graduating with the required human communication skills, in addition to being able to use AI effectively? And how will they ensure they are teaching and assessing those communication skills?

University assessments are changing

Universities are commencing an intensive period of assessment reform in response to AI. It will be a mammoth task. And it will require educators to think through the ways AI should be integrated into assessment tasks and teaching in the context of each discipline, and also when it should not be used. Considering both human communication skills and AI-mediated communication skills will be essential in this thought process.

Some universities are adopting a ‘two lane’ approach. In lane 1, assessments should be ‘secure’, meaning that the educator can be sure that learning outcomes have been met without the use of AI. In lane 2, use of AI is possible and often encouraged. Other universities are promoting a multi-lane approach to acknowledge the nuances between assessment types.

Regardless of the approach, many assessments – and definitely those in lane 1 – will  require structural changes, where the task is modified to ensure AI cannot be used.

These new assessments will therefore be assessing human interaction and communication skills in some form, be it written (such as handwritten notes), oral (interactive oral presentations) or multi-modal (oral check-ins on a written research report, or an oral presentation of a visual and written poster or design).

This shift is positive, as it also aligns with demand from employers and replicates authentic workplace tasks.

Assessing human communication skills

With human interaction and communication skills front and centre, they need to be explicitly taught and assessed. In the past, universities have not paid enough attention to communication skills, particularly oral and multimodal communication skills. Communication practices are even sometimes ‘invisible’ – meaning they are not explicitly included within the curriculum, or the teaching and assessment tasks.

Educators working at the subject or course level will need to ask:

What does effective human communication in this new restructured assessment (such as an interactive oral presentation) look like?

What verbal, visual and written communication skills do we want students to demonstrate in the task? For example: fluency, use of voice, verbal techniques to engage the audience, use of precise terminology, clear and engaging visuals, labelling and integration of visuals, coherent structure, and clarity of meaning.

How can these skills be incorporated into the assessment criteria and rubrics?

How can we scaffold the teaching of these skills over the term or semester to ensure ALL students have a good chance of performing well in the assessment?

Assessing AI-mediated communication skills

It is likely that many assessments will need to incorporate the use of AI to some degree. It may be that students use AI to edit a piece of writing, format a reference list, or help with brainstorming and researching a topic. Or AI could be used to complete a whole task and then compare it to a human-generated version. In any case, these tasks will require AI-mediated communication skills, which will need to be taught and assessed.

There has been much discussion about whether the traditional university essay is still an appropriate task in the age of AI. Some claim the essay is now ‘dead’. Others argue that writing is essential as a process to help us think, and this is often the case in the humanities. So rather than dismiss the take-home written essay or report task, perhaps it needs to be reframed – and taught and assessed – as a process that incorporates AI-mediated communication.

Educators working at the subject or course level will need to ask:

How can AI support students through the process of preparing a written, oral or multi-modal task in this subject or course?

How can students demonstrate their interaction with AI for the task? For example: through a declaration, a record of their interaction, or a guided reflection on their interaction.

Should students’ use of AI-mediated communication be assessed? If so, what does successful AI-mediated communication look like? For example: effective prompt writing, critical reflection on use of AI, careful editing and revising of AI text, synthesis of AI text with own voice.

How can these skills be incorporated into the assessment criteria and rubrics?

How can we scaffold the teaching of these skills over the term or semester to ensure ALL students have a good chance of performing well in the assessment?

Mapping communication skills

As part of university assessment reform, both human and AI-mediated communication skills should be mapped across degrees, incorporated into learning outcomes, assessment criteria and teaching activities.

UTS offers an example of a university-wide program to embed academic language that could be extended to embed human and AI-mediated communication skill milestones across degrees. An essential part of this mapping process will be facilitating effective collaboration between academic educators, language and communication practitioners and educational developers.

This article was written by Emily Edwards and Caroline Havery, an Associate Professor in the Teaching Learning and Curriculum Unit at the University Technology of Sydney.