What do you find to be most important about information security at JAE?

I believe that raising employee awareness around information security is extremely important. Therefore, we have formulated a mid-term plan and are currently working to roll it out across the company. Information security is handled differently around the world, and it's dependent on one's location and industry.

Europe is a step ahead of other countries in terms of GDPR compliance, while Asia is still in the early stages of implementation in some regions. While there are differences by region, we’re all protecting the same kind of data.

Additionally, with our employees now using cloud services, we are formulating internal management to control their use.

You’ve recently been involved in new initiatives, such as the introduction of chatbots and robotics processing automation (RPA), as well as the promotion of digital transformation (DX). Which technologies have been attracting your attention recently?

Generative AI is an interesting technology, depending on how it’s used. To properly use it in a professional setting, we must decide what kinds of regulations to establish. As someone who deals with security, I am concerned that it will be misused. It’s been said that if generative AI is abused, cyberattacks will take on a different dimension from what we’re seeing today.

On the other hand, I think it would be interesting to use AI to collect information from the internet and summarize it. This would be especially useful for summarizing articles. Here, there’s no need to input confidential company information and there are no security concerns—so I think it can lead to improved efficiency.

What are your company’s needs when using AI translation and machine translation?

• At JAE, our Sales and Marketing teams often need to communicate with overseas colleagues and clients, which can lead to difficulties when translating important emails and documents. They were using a free web translation service. However, it was difficult to translate content that included confidential internal information, and our teams were forced to replace sensitive text with different words —causing confusion.
• Also, our IT department sometimes needed to create documents in English when communicating with overseas partners. Although we have in-house translators, it was still difficult to translate documents into English in a timely manner. We did not want our communication to be hindered in this way, but we were unable to find a viable solution.

Thankfully, I heard about DeepL on a business trip to the UK when a local JAE employee told me that, "There is a service called DeepL that makes it easy to create documents in English."

Later, I spoke with our headquarters in Japan, I learned that some Japanese employees were already using the free version. Therefore, the Information Security Office decided to make DeepL Pro available to a wider range of employees across departments.

It must have been difficult to introduce software such as DeepL across the entire company. Do you have any advice for other companies that are considering using DeepL?

Deploying software across an entire company can be more difficult than necessary. In this situation, we chose to approach it from a security perspective. So, we made a company-wide announcement that, when translating confidential information or documents that must not be leaked outside the company, please use DeepL. At that time, we also created a document outlining how to use DeepL and the fees involved.

We also worked with IT managers in each country and company to deploy it, which was well received.

I think it would have been easier if we had first tried the service free of charge, and then expanded it to each company and department. It wouldn’t have been a bad approach to say, "This level of quality is available at no charge, and if you pay for it, in addition, security will be enhanced.” Recently, when we have proposed the use of translation software, we tend to focus on the advantages of using AI for translation. But I think it will be easier to convince people to use DeepL if we spotlight security.

When we hear about other companies, we get the impression that they have a hard time explaining return on investment (ROI) when making internal proposals or soliciting widespread use of the software. How do you calculate ROI?

Personally, I believe that DX and cloud services cannot be measured in terms of cost effectiveness. Since you are paying for centralized management, you need to be able to feel the effects of the service. I think it’s better to run internal surveys on a regular basis. However, it’s still difficult to measure the cost effectiveness of such a system.

Many people think that if you can't measure cost effectiveness, you shouldn't introduce a new system. But if you think that way, you will fall further and further behind other global companies.

Of course, it doesn’t mean we can invest without any limit, but if spending few thousand yen a month makes our work more efficient, we should invest.

What impact have you seen at JAE now that you’re using DeepL?

When I create English documents, I upload the Japanese documents as PowerPoint files to DeepL and translate them into English. We can now send the documents immediately without waiting for the translation.

In the past, this used to take a week to be done manually. DeepL has reduced the number of occurrences where we had to send materials but were too busy to do so. It has also improved the granularity of information sharing and overall communication.

Before I started using DeepL, it took me days to create a single English document. Since I started using DeepL, I can now do it in about a day.

Additionally, I think it has also improved my efficiency when translating emails from business partners, which often include confidential information, because I can translate them all at once. Our audit team has also been using DeepL for translating documents. Everyone reads and writes documents and emails in various places for different purposes, so it helps us save time.

How has JAE responded to the use of DeepL?

We have been using DeepL for about four months now, and our users have provided very positive feedback.

In terms of functionality, the document translation function is particularly useful, as we have had many requests to translate PDF files. I also often use PowerPoint to create English documents, so I feel that it has made my work more efficient and easier. Besides that, the ability to translate entire webpages with the Edge extension has been well received.

Alongside the quality, I also appreciate the fact that the company provides services on a global scale. From a security perspective, it is important to us that the company is headquartered in Europe and is GDPR compliant. In addition, since the service is clearly divided into American English and British English and covers a wide range of languages, it was easy to expand the service to other countries outside of Japan.

What would you like to work on in the future, not limited to the use of DeepL in particular?

As a member of the Information Security Department, I feel that the issue we need to address is how to understand and evaluate the security of the latest technologies. For example, we must decide whether or not to use generative AI and machine learning internally.

Of course, we believe that, first, we need to strengthen our own security measures. On top of 
that, we would also like to collect information on security in related areas and share it within the industry.