Japan’s New AI Code is one of the most interesting AI governance developments I’ve seen this year. A proposed “comply or explain” framework could hold AI companies accountable for training data transparency — without heavy-handed regulation. Here’s what it means for those of us buying or governing AI.
What Is Japan’s Proposed AI Code?
In August, a Japanese government expert panel broadly approved a draft code for businesses developing and providing generative AI. It’s non-binding, but it follows a “comply or explain” approach — companies either follow its principles or publicly explain why they haven’t. Japan is testing whether accountability can exist without fines or prescriptive rules.
This sits within Japan’s wider AI Promotion Act, enacted in 2025, which takes a pro-innovation stance while acknowledging that inappropriate AI use can contribute to copyright infringement and other harms.
The Three Core Principles
- Public disclosure: Developers and providers identify the models they use, provide an overview of training data types, how data were collected, and what internal controls protect intellectual property.
- Rights holder protections: Anyone who believes their work may have been used in training data can ask whether specific materials were included.
- User protections: Providers must respond to enquiries when a generated output raises a possible intellectual property issue.
Why “Comply or Explain” Is the Clever — and Difficult — Part
A voluntary code can still carry real consequences. If one AI company gives a meaningful account of its training practices while another offers only a vague excuse for not complying, buyers, public bodies, journalists and business partners can compare those responses. Reputational and commercial pressure can exist without a statutory penalty.
That said, disclosure doesn’t automatically equal good governance. Technology industry representatives argue that tracing individual sources is technically difficult for large models trained on enormous quantities of data. Those are real technical questions, even if they come from organisations representing industry interests.
But the opposite problem is equally real. Without enough information, a creator may suspect their work was used and have no practical way to test that suspicion. A public promise to respect copyright isn’t the same as evidence.
How This Differs From the EU’s Approach
For European listeners — this isn’t the same as the EU AI Act. The EU’s General-Purpose AI Code of Practice is voluntary, but it supports compliance with legal obligations that are not voluntary. Those include requirements to maintain a copyright policy and publish a sufficiently detailed summary of training content. Japan’s code relies more directly on public accountability and reputational pressure.
What This Means for Public Sector AI Buyers
My takeaway is simple: don’t wait for every rule to be finalised. If you’re buying or governing AI, start asking your suppliers these questions now:
- Which underlying models do you use?
- What can you tell us about the categories and sources of your training data?
- What protections do you have for intellectual property?
- How can a rights holder or user challenge a potentially infringing output?
If they can’t answer, ask them to explain precisely why — and make sure that explanation is specific and evidence-based, not a carefully worded corporate paragraph.
If you want to build your AI literacy and skills, I can help
📚 Grab AI Playbook on Amazon, the perfect podcast companion.
📰 Subscribe to Human in the Loop, my LinkedIn Newsletter
🎓 Take the CPD Accredited AI Playbook Diploma
📞 Book a call at digitaltraining.ie
Key Statistics
– Japan’s AI Promotion Act was enacted in 2025.
– The draft AI code was broadly approved by a Japanese government expert panel in August.
– The code focuses on two core issues: intellectual property protection and training data transparency.
– The “comply or explain” mechanism is non-binding — companies either comply or publicly explain why they have not.
About the Podcast Host
AI and Communications Specialist | Author | Podcaster | International Speaker
Joanne Sweeney is a recognised leader in digital marketing and digital transformation, specialising in the practical application of AI to enhance communication strategies. The founder of Digital Training Institute, Joanne, empowers leaders and communications professionals to leverage AI responsibly and effectively, ensuring transparency, trust, and tangible results.
She completed the Harvard Kennedy School Executive Programme in AI and Public Policy this summer before launching her fourth book AI Playbook, in September. Joanne has also published two editions of Public Sector Marketing Pro, which were preceded by her first title, Social Media Under Investigation Law Enforcement and the Social Web.
Her books have become go-to resources for governments, public sector and policing agencies, as well as marketing and communications professionals, all navigating the challenges of the changing Internet.
Joanne is the host of the AI SIX Podcast, a daily six-minute podcast that provides actionable insights, expert interviews, and practical advice to help professionals navigate the new world of work.
