Abstract This research investigates the post-editing (PE) of machine- and artificial intelligence (AI)-generated technical documents in the context of tight deadline jobs (TDJs) to assess team dynamics and efficiency. In a mock translation agency setting, 12 trainee translators were divided into two groups: Team 1 focused on PE Baidu’s machine translation (MT), while Team 2 worked on the outputs from Kimi AI, following a convergent parallel mixed-methods design. Quantitative metrics monitored time, errors, and consulted resources, while qualitative insights explored stress levels and teamwork. Results indicate that Team 1 invested more time (60 vs. 48 minutes) and corrected more errors (19 vs. 12) due to the prevalence of mistakes in Baidu’s output. They also consulted more resources but achieved lower accuracy (92% vs. 95%) than Team 2, which improved Kimi AI’s fluent yet contextually flawed translations. Both teams tended to prioritize speed over quality, with ongoing challenges in terminology. The study's implications for the professional approach to translator training (PATT) involve simulating AI/MT PE to improve error detection, making contextual adjustments, incorporating stress management strategies, and promoting terminology proficiency and flexible team roles. This research helps address gaps in PATT related to TDJ training, machine-translated post-editing, AI functionality integration, and specialized text training, thereby enhancing translator education to align with current industry needs.
Honghui Hu (Thu,) studied this question.