This research investigated the effect of native and foreign languages and responsibility attribution to self and AI during collaboration. The study performed a 2 (language: Chinese, English) × 2 (outcome: success, failure) mixed-design experiment, with Chinese as their native language and English as their foreign language. During the experiment, changes in blood flow within the prefrontal cortex were monitored using a functional Near-Infrared Spectroscopy (fNIRS). Tasks conducted in native tongue were approached with greater ease and confidence, which was reflected in their decision-making patterns; individuals were less likely to be swayed by different opinions or choices. In native language scenarios, there was a tendency to attribute success to oneself, stemming from a sense of confidence and control, supported by the self-serving bias. The case was opposite in a foreign language context, possibly due to reduced self-confidence in language proficiency. Failures were more readily attributed to AI when operating in a native language, possibly as a means of deflecting personal blame, whereas in a foreign language, the attribution of failure swung back towards oneself, reflecting uncertainties in handling the language effectively.