Latest

LatestRegulation

Japan's AI legal-system working group meets again, with penalties reported as under discussion

The Cabinet Office page for the second meeting of the working group on the shape of the AI legal system lists the meeting as 11 November 2026, from 10:00 to 12:00, at Akasaka Green Cross. The meetings index still marks the minutes as in preparation. The page links a paper titled as the day's items for discussion. That paper asks the group to discuss which risks need a further response, how to draw a line if AI above a set level of capability is regulated, who would be covered if hard law is used, what duties and state powers that would involve, what role the AI Safety Institute should have, and what further steps belong under existing laws and guidelines. A table in the paper treats AI-specific hard law for higher-capability systems and specified risks as the cell to discuss, phrased as whether that need is high. The other cells say existing laws and guidelines. The paper is a discussion agenda. It is not an amendment.

Source: Cabinet Office of Japan (opens in a new tab), 11 Nov 2026Also: Kyodo News, 12 Nov 2026 (opens in a new tab)

The Cabinet Office page for the second meeting of the working group on the shape of the AI legal system lists the meeting as 11 November 2026, from 10:00 to 12:00, at Akasaka Green Cross. The meetings index still marks the minutes as in preparation. The page links a paper titled as the day's items for discussion. That paper asks the group to discuss which risks need a further response, how to draw a line if AI above a set level of capability is regulated, who would be covered if hard law is used, what duties and state powers that would involve, what role the AI Safety Institute should have, and what further steps belong under existing laws and guidelines. A table in the paper treats AI-specific hard law for higher-capability systems and specified risks as the cell to discuss, phrased as whether that need is high. The other cells say existing laws and guidelines. The paper is a discussion agenda. It is not an amendment.

Kyodo News, in an article timed 12 November 2026 at 10:30, reports that a source familiar with the matter said on Wednesday that the government is considering requiring AI-related businesses to take safety measures, with penalties for violations. As reported by Kyodo, the approach under discussion would let the government issue legally binding orders to businesses that do not cooperate after an incident, with penalties expected to be limited to incidents involving advanced AI, and the working group aims to present an outline of its review, including possible revisions, by year-end. Kyodo says the current AI law, enacted in May 2025, has no penalty provisions and that its measures are nonbinding. The Cabinet Office paper opened for this account does not itself set a penalty or a year-end date.

What it means for business: the AI Promotion Act is unchanged. No new duty and no penalty are in force. The 11 November meeting is a discussion of whether hard law should cover higher-capability AI. The penalty and year-end lines are Kyodo's report of a source, not a published bill. This describes those pages. It is not advice on a system used in Japan.

Get the free AIIntel Briefing, daily or weekly

All items in Regulation