What the US-China AI Dialogue Means for SME Governance

What the US-China Alignment on AI Governance Means

In May 2025, it was reported that the US and Chinese governments reached an agreement during intergovernmental talks on AI (artificial intelligence). China’s Ministry of Foreign Affairs announced it would “promote governance,” and major media outlets like Asahi Shimbun covered the development.

It would be premature to dismiss this news as something happening in a distant world. As international rule-making for AI accelerates, it will indirectly impact the cloud services and business systems used by small and medium-sized enterprises (SMEs).

For example, suppose your company analyzes customer data using AI. The governance standards agreed upon by the US and China could very well be reflected in domestic regulations through the Japanese government.

At that point, “I didn’t know” won’t be an excuse. As a business owner, it’s time to start thinking about how to design your company’s AI usage.

Where Should SMEs Start with AI Governance?

Large corporations have dedicated AI ethics committees and specialized teams. SMEs don’t have those same resources. So, what should you do?

The answer is simple: the business owner themselves must decide “what AI is allowed to do and what it is not.”

Specifically, keeping the following three points in mind can help you avoid major risks.

1. Take Stock of Tasks That Rely on AI Decisions

Are there any tasks in your current operations where you use AI output as-is?

For example, cases where AI handles the initial screening of job applicants, or where AI scores determine credit decisions for business partners.

For such tasks, you need a system to regularly check whether AI decisions contain bias.

For an SME, even having the business owner directly check a sample once a month can be effective. Rather than aiming for a perfect system, the key is to first build the habit of “looking.”

2. Formalize Rules for Handling Customer Data

Whether the data used for AI training is appropriate is also a core aspect of governance.

Especially when inputting personal or confidential information into an AI service, you need to verify how that data will be handled.

Specifically, read the terms of service for the AI services you use and check whether your data will be used for training. If it’s unclear, contact the provider to get clarification.

Doing this work just once can significantly reduce the risk of “customer personal information being leaked externally” later on.

3. Intentionally Leave Decision-Making Areas for Humans

AI is often thought to be all-powerful, but that’s not the case.

In particular, the following types of decisions should be made by humans:

• Employee evaluations and promotions
• Final decisions on contract terms with business partners
• Direction of management strategy

AI is merely a “tool for organizing information.” It’s crucial to share the principle within your company that final decisions are made by humans.

Bridging International Rules and Your Company’s Rules

What the US-China AI dialogue shows is that we’ve entered an era where governance advances through both “rules between nations” and “rules on the ground.”

SME business owners need a sense of balance to catch international trends while applying them to their own companies.

Specifically, it’s a good idea to develop a habit of regularly checking the following information sources:

• AI-related guidelines from Japan’s Ministry of Economy, Trade and Industry (METI)
• Materials for SMEs from the Japan Deep Learning Association
• Guidelines on AI utilization from industry associations

However, don’t just stop at gathering information. The most important thing is to decide on “rules for using AI” tailored to your own business.

Three Points to Avoid Failure

Here are some common failure patterns.

Failure 1: Making AI Adoption the Goal Itself

This is when a company becomes satisfied with the fact that it “introduced AI” and neglects subsequent rule-making. AI is a means, not an end. Without governance design after implementation, you only increase risk.

Failure 2: Not Acting Because You Seek Perfection

This is thinking, “We don’t have an AI governance expert, so it’s impossible,” and doing nothing. SMEs don’t need perfect governance. The realistic goal isn’t to reduce risk to zero, but to keep it within an acceptable range.

Failure 3: The Business Owner Not Being Involved

This is leaving AI utilization entirely to the staff on the ground. Since teams prioritize efficiency, they may use AI in high-risk ways. A system is needed where the business owner regularly checks the situation and adjusts policies as necessary.

Actions You Can Take Starting Today

Finally, here are three actions you can put into practice immediately.

1. At next week’s morning meeting, discuss with your team “which tasks AI can handle and which it should not.”
2. Re-read the terms of service for the AI services you use by the end of this month.
3. Set aside time once a month for the business owner to directly check AI output results.

None of these require additional costs or personnel. They can be achieved with just the business owner’s awareness and a little bit of time.

International discussions on AI governance will only accelerate from here. To avoid being left behind, take that first small step today.

We hope AI utilization in your company becomes an engine for growth, not a source of risk.