How “SynthID” AI Identification Technology is Transforming Business Risk Management

How to Prepare for New Management Risks in the AI Era

As generative AI becomes more widespread, a new challenge is emerging for business leaders: identifying AI-generated content.

Recently, South Korean tech giant Kakao joined the list of companies adopting “SynthID,” an AI content identification technology developed by Google. This news goes beyond a simple tech rollout, prompting a rethink of how to manage risks in the age of AI.

As someone who has leveraged AI to achieve a reduction of 1,550 work hours annually in my own company, I deeply understand the importance of this technology. The more we use AI, the greater the need to properly manage which content is AI-generated.

What is SynthID and Why Should Business Leaders Care?

SynthID is a technology developed by Google that embeds digital watermarks into AI-generated content. These watermarks are invisible to the human eye but can be detected with specialized tools. Kakao’s adoption stems from the need to mitigate risks of AI-generated content spreading across its platforms like KakaoTalk and KakaoPay.

From a management perspective, the key is that this technology directly ties into “compliance” and “brand value protection.”

For example, if your company mass-produces marketing materials using AI, what happens if that content inadvertently learns from copyrighted works of others? Or what if fake reviews or misinformation spread through your services? These risks become more tangible as AI use becomes routine.

Identification technologies like SynthID provide the foundation for visualizing and managing these risks.

Implementation Costs and Practical Applications

Naturally, cost is a concern. Currently, SynthID is offered as a Google Cloud service, with usage-based pricing depending on the volume of content processed. Large-scale implementations could cost tens of thousands of yen (hundreds of USD) per month, but entry-level plans are available for small and medium-sized businesses to try out affordably.

My recommendation is to start with an “experimental rollout.” Here are some steps to consider:

First, identify your company’s AI-generated content. List everything created by AI, such as internal blog posts, social media updates, and customer emails.

Second, apply SynthID to these items for management. This way, even if content leaks or is misused, you can prove it was AI-generated by your company.

Third, establish a verification system for incoming AI-generated content from external sources. For instance, create a rule to check if proposals or design data from business partners have SynthID watermarks.

Learning from Sumitomo Life’s Case: Balancing AI Use and Risk Management

Recently, Sumitomo Life Insurance also made news for using AI to check for missed insurance claims. The goal was to improve efficiency for office staff, with AI analyzing past claim data to automatically detect potential omissions.

This case perfectly illustrates the essence of AI use. AI should be used to “complement human judgment,” not “replace it.” At Sumitomo Life, human staff make the final call on the “potential omissions” flagged by AI.

Here again, the “reliability” of AI-generated analysis results is questioned. If the AI makes an error, it could lead to missed or overpaid claims. Technologies like SynthID make AI outputs traceable, simplifying audits and corrections.

Three Actions Business Leaders Should Take Now

So, what should CEOs and CTOs do specifically? Based on my experience, here are three high-priority actions.

Conduct an “Inventory” of AI-Generated Content

First, identify all content generated by AI within your company. You’ll likely be surprised at how many areas use AI, including blogs, newsletters, social media posts, customer support replies, and contract drafts.

In my own company, we use AI across 29 business areas, including automated social media posting, contract review, and manual generation. We’ve implemented a system to tag each output. This alone prevents confusion later about whether content was AI-generated.

Establish Rules for Accepting External AI-Generated Content

Content from partners or freelancers may also include AI-generated material. For example, if design assets or articles from an outsourced vendor are AI-generated, there’s a risk of copyright infringement.

Consider adding a clause to contracts or purchase orders stating: “If AI-generated content is used, it must be disclosed and identification technology must be applied.” This is something I actually recommend in contract reviews for clients.

Start a Trial of AI Identification Technology

SynthID isn’t the only option; several technologies can identify AI-generated content. Start with free trials or small-scale tests to see if they fit your workflow.

Cost-wise, some services start at a few tens of thousands of yen (a few hundred USD) per month. Given the potential risks—like copyright lawsuits or reputational damage costing millions of yen (tens of thousands of USD) annually—the return on investment is well worth it.

Summary: The Next Phase of AI Use is Management and Trust

With generative AI adoption rates reaching 46.5%, the focus for business leaders is shifting from “how to use it” to “how to manage it.” Kakao’s adoption of SynthID is a symbolic milestone.

AI is undoubtedly a powerful tool. But if not handled correctly, it can harm your own company. Identification technology acts as a “safety device” for advancing AI use securely.

As a leader, you don’t need to slow down AI adoption. However, you must simultaneously build risk management systems. This balance is key for companies to thrive in the AI era.

Take the first step today in managing your company’s AI content. That decision will significantly impact your company’s value in the years to come.