The city of Yokosuka in Japan’s Kanagawa Prefecture has officially integrated the artificial intelligence chatbot, ChatGPT, into its administrative operations. This comes in the wake of a successful month-long trial during which the AI tool demonstrated its ability to significantly improve operational efficiency and reduce business hours.
Pioneering AI Use in Government Operations
Yokosuka is the first local government in Japan to test the generative AI, ChatGPT. Built upon a machine learning model that functions similarly to the human brain, the AI tool proved to be instrumental in numerous administrative tasks during the trial period. City officials found it particularly useful for:
- Drafting bulletins
- Summarizing meeting records
- Proofreading documents for typographical errors.
Prospects of Continued Use
The municipal government estimated that the continued use of ChatGPT could trim daily working hours by roughly 10 minutes, a boon in a country known for its long working hours. The Mayor of Yokosuka, Katsuaki Kamiji, endorsed the AI tool as a valuable asset that could potentially curtail labor costs, and underscored the importance of integrating AI in clerical work.
Despite the promising results, the trial did reveal that about half of the city staff surveyed were unsatisfied with the chatbot’s response accuracy. In response to these concerns, the city intends to solicit expert advice and organize competitions to develop effective prompts that will ensure more precise responses from ChatGPT.
Security and Privacy Issues
Like all AI technology, the use of ChatGPT has raised questions about data security and privacy. In order to assuage these concerns, the Yokosuka government has forbidden its officials from entering personal data into the chatbot. Furthermore, Japan’s privacy watchdog issued administrative guidance to OpenAI, the U.S. developer of ChatGPT, last week, advising it not to collect sensitive private data without the individual’s consent. The Personal Information Protection Commission also alerted administrative agencies to the privacy risks posed by entering personal information into the chatbot.