Artificial Intelligence in Applied Cognitive Psychology: A Commentary

Le résumé fourni par la source

AI integration in applied cognitive psychology demands critical evaluation beyond efficiency metrics. Despite widespread institutional adoption, emerging research reveals concerning patterns including high hallucination rates, deteriorating retention with prolonged exposure, and a consistent tendency to support surface-level task completion at the expense of deeper cognitive processing. These findings align with established principles regarding desirable difficulties, metacognitive monitoring, skill acquisition, and vigilance, suggesting that applications prioritising task completion over cognitive development risk undermining the adaptive expertise essential for complex professional contexts. Methodological weaknesses in existing research, including brief interventions, inadequate control comparisons, and reliance on satisfaction measures, further constrain confident conclusions. Nonetheless, several domains including cognitive accessibility, rehabilitation, vigilance, and adaptive tutoring represent areas of genuine promise where AI’s architecture may complement rather than conflict with established cognitive science. This commentary synthesises emerging evidence, examines methodological limitations, proposes research priorities for responsible integration, and reflects on where cautious optimism is warranted.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.