Una valutazione col pilota automatico? Una riflessione sulle cose che possiamo guadagnare e quelle che rischiamo di perdere impiegando l’intelligenza artificiale nei processi valutativi

Abstract

EVALUATION ON AUTOPILOT? A REFLECTION ON THE THINGS WE CAN GAIN AND THOSE WE RISK LOSING BY USING ARTIFICIAL INTELLIGENCE IN EVALUATION PROCESSES

Abstract

The paper analyses pros and cons of using generative artificial intelligence in the students’ learning assessment process. In particular, the article focuses on the possibility offered by programs such as ChatGPT to reduce the time devoted to complex operations such as construction of multiple-choice questions or classification of open-ended answers. Thus, it is possible to use the generated content to support a phase of the evaluation process that normally dilutes time and makes the workload difficult to sustain. Although the benefits from this side are significant, the contribution emphasises that it is quite inappropriate to claim that generative artificial intelligence can take over the responsibility for judgement processing. Firstly, since there are no objective evaluations, such a choice would merely conceal the necessarily subjective criteria used to make well-founded value judgements. Secondly, such a choice would imply a deresponsibility with respect to judgements that have relevant practical consequences and a renunciation of the possibility of actively and consciously employing the evaluation process for the enrichment of future experiences.