Discussion

Japan trials generative AI to draft parliamentary answers

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#10572

Japan’s Digital Agency is trialling a generative AI app to help government workers prepare answers to lawmakers’ questions. It can analyse a question and produce several draft responses in minutes, giving officials a new tool for a central part of parliamentary work.

Watch Desk analysis

What happened

The app runs on the agency’s Gennai platform, which is available to about 180,000 government workers. Officials can enter a lawmaker’s question; the system identifies its key points and generates draft answers using past government responses and records of parliamentary deliberations. It can also help officials understand the question’s intent in advance.

The app offers different writing modes, including one for explaining policy direction and another for reflecting political considerations. The agency plans to assess its effectiveness during the current Diet session, with full-scale operation targeted for next year’s ordinary session. Jiji Press reported the trial on 2 November.

Why it matters

This is a concrete use of generative AI inside government, where speed is useful but the wording of an official answer matters. Drafting several options quickly could give officials more time to check the substance, or make it easier for polished text to outrun careful scrutiny. The report describes assistance with drafting, not AI making or approving the government’s answers.

The trial’s evaluation will matter more than the existence of the app. Officials will need to judge whether it saves time without muddling the question, repeating outdated answers or blurring who is responsible for the final response.

Our read

This is a meaningful test of AI in the machinery of government, not another chatbot staged beside a flag. The useful detail is that the system draws on past answers and parliamentary records, while offering multiple drafts and distinct writing modes. Whether that helps officials respond more clearly is still to be tested. The agency should publish what it measures and how human review works before calling the trial a success.

What to watch

  • What the Digital Agency measures during the current session, including accuracy, time saved and officials’ workload.
  • Whether the system’s drafts are checked against current policy and source records.
  • Whether full-scale use begins next year, and what safeguards or review process accompany it. Activity teaser: Japan’s Digital Agency is testing generative AI that analyses lawmakers’ questions and drafts several possible government answers in minutes, using past responses and parliamentary records. The app also offers different writing modes, from policy explanation to wording that reflects political considerations. The agency plans to evaluate the tool during the current Diet session before deciding whether to move towards full-scale use next year. Faster drafting is useful; getting the answer right, and keeping a person responsible for it, is the real test.

Discussion spark: Should AI-drafted parliamentary answers be published with a clear record of human review, or is responsibility for the final answer enough?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.