Task-based language teaching views tasks as devices for instigating the kinds of input, output and interaction that facilitate language acquisition. Researchers have approached tasks from this perspective in three ways.
Early research consisted of “task-based learner performance studies” Plonsky and Kim (2016: 74). These are studies that manipulated various dimensions of task design and implementation conditions in order to investigate what effect they have on interaction and were informed by the Interaction Hypothesis (Long, 1983; 1996) and the importance of the negotiation of meaning. This research was successful in identifying a number of task design and implementation features most likely to result in negotiation sequences (e..g. two-way tasks that required information and tasks that a closed outcome).
Later research involved tasks-as-treatment studies, where the effect of different task design and implementation variables on learner production was investigated using measures of complexity, accuracy and fluency.
In the third approach, researchers have investigated tasks in terms of learner engagement. Engagement is operationalized in terms of four aspects – behavioural, cognitive, affective and social. This research was motivated by the claim that unless learners engage with a task, no learning can take place.
How AI can be used to evaluate L2 learners performance of monologic and dialogic tasks will be illustrated.