On 20 October 2026, Japan's Cabinet decided to establish the AI and Digital Reform Promotion Council. The Digital Agency's council page says the body will promote cross-government work on reviewing economic and social regulations, systems and operational rules, and on implementing AI transformation (AX) and digital transformation (DX) in society. The Digital Policy Alert record of the decision dates it to the same day. The question for this publication is whether the council becomes a deregulatory engine or another layer of coordination.
What the council is
According to the Digital Agency's council page, the council replaces the Digital Administrative and Fiscal Reform Council, which was set up by Cabinet decision on 6 October 2023 and is now abolished. Matters that body was discussing carry over. The Digital Agency acts as secretariat, supported by the Cabinet Secretariat and Cabinet Office. Its members are the relevant ministers, including those for internal affairs and communications, finance, and economy, trade and industry. Ten outside experts also sit on it, drawn from academia, industry, and local government. Papers are published as a general rule, and minutes are public except for portions the chair withholds.
That transparency rule matters. Reform bodies often work through informal channels, and a published paper trail lets industry and civil society see which rules are on the table.
The first meeting
The council met for the first time on 22 October 2026 at the Prime Minister's Official Residence. According to the Digital Agency's record of the meeting, the session ran from 16:20 to 16:50. The agenda covered how the council will operate, proposed discussion points, and an exchange of opinions. The minister said the council would compile three things by the end of the year: a "Specific Image of AX," the "Principles for AI and Digital Reforms," and an "Action Plan to Accelerate Social Implementation." The record also says the AX vision would feed into the revision of the AI Basic Plan. A Working Group on Institutional Reform will sit under the Task Force on AI Strategies, and revising legal systems including the ID Law is under consideration.
The meeting followed instructions from Prime Minister Takaichi, and an interim report is due at year-end. The council has about two months to produce principles that will shape the next AI plan.
The strongest case for a heavier hand
The case for more binding rules is serious. Japan's AI governance rests on a promotion statute with no monetary penalties. When the Cabinet revised the AI Basic Plan in July 2026, Yomiuri Shimbun reported through the Asia News Network that the government said it would proactively and continuously review the system, including the AI law, to deal with risks more effectively. Possible measures include penalties for operators who refuse to cooperate with investigations. Advocates argue that soft law leaves citizens exposed when a powerful system fails, and that a law with no enforcement tools cannot discipline large vendors.
That argument has force where a clear harm exists and an existing law cannot reach it. It is weaker as a general case for new mandates ahead of evidence.
Why the deregulatory framing is the right one
The council's stated mandate is reviewing regulations, systems and operational rules, not writing new obligations. That is the right emphasis. Japan's AI adoption problem is mostly friction in the old rulebook: paper-era procedures, siloed ministries, and ID and data-sharing rules that predate cloud computing. The mention of ID Law revision shows the council is looking at that kind of plumbing. A reform agenda that removes those frictions does more for AI use in public services and business than a new compliance regime would.
Japan has also shown it can regulate digital markets without a sweeping horizontal AI law. Its Mobile Software Competition Act, which PPC Land reports took full effect in December 2025, applies to three designated entities: Apple Inc., iTunes K.K. and Google LLC. The designation came in March 2025. The law permits restrictions only on five closed grounds, including cybersecurity, user information protection and protection of minors. This is a targeted, named-company approach. It is a better template than a rule that covers every developer.
What to watch before year-end
- Principles that bind the government first. The strongest outcome would be that the "Principles for AI and Digital Reforms" apply to ministries' own procurement and service delivery, with measurable deadlines, rather than to the private sector.
- Sunset and evidence tests. Any proposal for new obligations should state the specific harm, show why existing law is insufficient, and include a review date. The July revision's language about reviewing the AI law should be tested against that standard.
- The working group's remit. The Working Group on Institutional Reform should publish which statutes it is reviewing. If the list is mostly legacy procedural rules, the council is on track. If it is mostly new duties on model developers, the framing has drifted.
- Handling of the ID Law. Revisions that make identity verification interoperable are useful, but they raise data-protection questions that deserve open consultation rather than a short interim-report cycle.
Conclusion
The council's design is sensible: it has ministerial authority, outside experts, a public paper trail and a deadline. The risk lies in the output. A year-end set of principles that says little about which rules will be removed would add another layer of coordination, and one that quietly authorises penalties would break from the light-touch approach that has made Japan an attractive place to deploy AI. The council has produced no substantive decisions yet, so judgment should wait for the interim report. The standard to hold it to is simple: every proposed rule should come with named evidence of harm and a review date.