Tokyo [Japan], June 7: The Japanese city of Yokosuka has officially applied ChatGPT to administrative activities since June 5, after a month of testing showed that this artificial intelligence (AI) tool helps improve work efficiency.
Yokosuka City, in Kanagawa Prefecture south of Tokyo, has become the first local government in Japan to test experimental ChatGPT , a Generative AI application driven by an active machine learning model. like the human brain.
During testing, city officials used ChatGPT to create newsletters, summarize meeting content, and correct typographical errors in documents, among other purposes, the company reports. believe in Kyodo.
The city government estimates that if ChatGPT continues to use, the working time can be reduced by at least 10 minutes per day.
"We have a great tool. It will help reduce labor costs. Putting AI in clerical work is important," Mayor KatsuakiKamiji said at a recent press conference.
But the city also noted that about half of those surveyed said they were dissatisfied with the accuracy of ChatGPT's responses.
To upskill, the city will seek advice from experts and hold case competitions on how to create effective "prompts" so that a chatbot like ChatGPT gives a more accurate response.
Chatbots are software applications that are trained using vast amounts of data from the internet, allowing them to process and simulate human-like conversations with users. They provide answers based on the user's instructions and questions.
There are concerns about confidential information being leaked as a result of using ChatGPT. The Yokosuka government said it does not allow city officials to enter personal information when using the app.
Last week, Japan's privacy watchdog issued an administrative directive to OpenAI , the US company behind ChatGPT, asking the company not to collect sensitive personal information, such as medical records or criminal record, without the consent of the individual.
The Japanese government's Committee for the Protection of Personal Information also warned administrative agencies about the risk of privacy invasion stemming from entering personal information into chatbots.
Source: ThanhNien Newspaper
