生成式AI科技執法意願之初探

ChatGPT自2022年11月底推出3.5版後，短短不到2個月的時間用戶累計超過1億人。由於其可與人類進行複雜對話、寫報告、寫程式和通過專業考試等，讓人們爭相成為會員下載使用，且留下深刻體驗。然而我們對於人們使用ChatGPT等生成式AI的動機與可能帶來的知覺利益與風險之理解至今仍然有限，特別是需用AI科技執法的警察人員。為解決此一實務問題，本文使用行為推理理論進行延伸建構實證模型，然後以偏最小平方結構方程模型（PLS-SEM），進行生成式AI應用程式的使用動機與阻礙因素之實務數據檢測與結果驗證。為測試警察人員生成式AI使用意願模型，我們採便利抽樣，共取得157份有效參與者資料。調查結果表明，科技樂觀、創新先驅、有用知覺、趨勢知覺與智慧知覺會顯著影響員警生成式AI科技執法的使用態度。此外，科技樂觀、科技不安、易用知覺、趨勢知覺與使用態度會直接對於AI科技執法的員警意願扮演關鍵作用。因此，本文可補充相關文獻的不足與提供實務建議。

關鍵字

生成式AI ；行為推理理論；科技準備；產業創新； ChatGPT

並列摘要

Since ChatGPT launched version 3.5 at the end of November 2022, the number of users has exceeded 100 million in less than 2 months. Because it can finish a lot of tasks in daily life and work like human being, people rush to be membership. However, our understanding of people's motivations, the perceived benefits and risks for using generative AI such as ChatGPT are still vague, especially for police officers who need to use AI technology to do their job well. For solving this problem, this study based on behavioral reasoning theory to construct an empirical model and use PLS-SEM to test hypotheses. We adopted convenience sampling and obtained a total of 157 valid participants' data. The results show that optimism, innovativeness, perceived usefulness, perceived trendiness and perceived informativeness will significantly affect the attitude of police officers towards the use of generative AI technology. In addition, optimism, insecurity, perceived ease of use, perceived trendiness and attitude will directly play important roles in the police's intention to use generative AI technology. Therefore, this article can supplement the shortcomings of relevant literature and provide practical suggestions.