This study aimed to explore the cognitive mechanics in AI-based games and ‎provide a preliminary evaluation of their impact on the components of creativity ‎‎(fluency, originality, flexibility, and elaboration) within an exploratory framework. ‎The present mixed-methods research utilized an exploratory single-group quasi-‎experimental design (pre-test/post-test). In the qualitative phase, through a ‎systematic content analysis of six selected games equipped with adaptive and ‎generative algorithms, key mechanics such as "interactive co-creation" and ‎‎"dynamic challenge adaptation" were identified. In the exploratory quantitative ‎phase, 12 children aged 4 to 7 in Tehran participated in a 4-week intervention ‎program (three 45-minute supervised sessions per week), selected through ‎purposive sampling with informed parental consent. The data collection tool was ‎the pictorial version of the Torrance Test of Creative Thinking (Form A), for which ‎inter-rater reliability was calculated at 0.87. Quantitative findings indicated a ‎significant increase in the total mean creativity scores from pre-test (M=85.50, ‎SD=5.20) to post-test (M=112.40, SD=6.10). The results of the paired t-test ‎revealed a statistically significant difference (t(11)=21.42, p<0.0001). The ‎calculated effect size (Cohen’s d=6.18) was very large. However, given the ‎exploratory nature of the study, the lack of a control group, and the small sample ‎size, these findings should be considered as preliminary evidence of the potential ‎of intelligent games to facilitate creativity, rather than conclusive proof of a causal ‎relationship. Qualitative analysis suggests that AI, by providing a platform for "safe ‎trial and error" and personalizing challenges, may act as a cognitive facilitator.‎