
Photo: Wanan Yossingkum | Dreamstime.com
Japanese city trials ChatGPT
24 April 2023
by Sarah Wray
The city of Yokosuka in Japan has launched a trial of OpenAI’s advanced chatbot ChatGPT across all departments.
ChatGPT’s API has been linked with the city’s business collaboration tool LoGo Chat.
The city said all employees could use it to write drafts, summarise information and meeting minutes, check for spelling mistakes and even “create ideas”.
A statement said: “This is expected to improve the efficiency of operations, and we expect that the wide use [by] staff will lead to the creation of various use cases.”
Managing risk
The city said it will ensure that the information entered into ChatGPT is not used for secondary purposes, and that confidential information and personal information are not handled.
Yokosuka’s smart city and digital government policies promote using technology to carry out operations efficiently and effectively.
“As a result, staff will focus on work that can only be done by people,” the city said.
The experiment is already underway. The news announcement read: “This release was drafted using ChatGPT, and staff proofread it.”
City experiments
Other governments including Dubai and Singapore have also outlined plans to use ChatGPT – or in Singapore’s case, the large language models that underpin it.
Since ChatGPT burst onto the scene in November, followed by Google’s Bard last month, they have spurred widespread debate over the opportunities and risks presented by generative artificial intelligence. The latter includes potential inaccuracies and misuse by bad actors, as well as the impact on jobs and education.
Earlier this month, a regional Australian mayor said he may sue OpenAI for defamation if it did not correct false claims that ChatGPT shared about him.
ln March, Italy announced a temporary ban on ChatGPT due to privacy concerns. OpenAI has said it complies with privacy laws.
The leader of Italy’s data protection watchdog since said Italy would be ready to allow the return of ChatGPT if OpenAI takes “useful steps” to address the concerns.