[NEC x Anthropic] What are the 'Four Sovereignties' President Morita sought to protect? Dissecting corporate defense strategies in the AI era🛡️

Hello everyone. This is Kojima.

Today is July 16, 2026. The evolution of AI is truly dizzying, isn't it?

To all the engineers out there complaining, 'Ugh, the prompt that worked yesterday is throwing errors today...'—thank you for your hard work again today!😅

Now, there is news that has been causing quite a stir in the IT industry recently.

It is that the Japanese company NEC formed a partnership with the major US AI firm Anthropic in just three weeks.

It's so lightning-fast that you might wonder, 'Isn't three weeks shorter than the time it takes for internal approval documents to circulate?'

However, there is an important point here.

It seems that NEC President Takayuki Morita was thoroughly committed to the 'four sovereignties' throughout this partnership.

What kind of sovereignties are they exactly?

And why is this becoming a life-or-death issue right now?

Today, I will delve into the technical abyss behind this and the points we engineers should be mindful of starting tomorrow!

Surprising fact: The 'sense of crisis' behind the three-week partnership

First, there is a shocking figure.

The time it took for the partnership between NEC and Anthropic to be established is said to be only three weeks.

Normally, it wouldn't be strange for a partnership between large corporations to take six months to a year.

Why was it at such a rapid pace?

It lies in the tremendous capabilities of Anthropic's latest model, 'Mythos'.

Mythos is attracting attention not only for code generation but also for its ability to oversee the entire system architecture and discover points like 'this is vulnerable' or 'there is a bug in this logic'.

This is convenient, but on the flip side, it also means 'you can use AI to find ways to break systems from an attacker's perspective'.

Governments and security agencies around the world are deeply concerned about what would happen if powerful analytical capabilities like those of Mythos were misused.

In other words, NEC's rapid movement was not just about securing access to this cutting-edge AI technology.

In an era where "AI-driven attacks" are on the rise, it is believed they saw a path to victory in the construction of a defense line that ensures how to protect their own systems and guarantee reliability.

What are the "four sovereignties" that President Morita is so particular about?

Here, I will touch upon the main topic of the "four sovereignties" (sovereignty).

This is not just a slogan; it is no exaggeration to say that these are the conditions for corporate survival in the AI era.

Specifically, they are the following four:

-Data sovereignty: The right to manage where your company's important data is stored and who can see it

-Algorithm sovereignty: The right to understand and control the behavior and biases of the AI models you adopt

-Infrastructure sovereignty: The right to manage your dependence on the foundation (cloud or hardware) that runs the AI

-Decision-making sovereignty: The right for humans to hold final responsibility and decision-making power regarding judgments made by AI

It seems there is a movement emerging to collectively call these "Sovereign AI."

Let's understand with an analogy: Eating out vs. Home cooking

This may be a difficult concept, so let's compare it to cooking.

Conventional cloud services and the use of black-box AI are like a "high-end restaurant (eating out)."

If it tastes good, there are no complaints, but in many cases, you cannot ask in detail what it is made of or about the cooking process, right?

If there were a problem with the ingredients, how would you hold them accountable?

On the other hand, a state that possesses the "four sovereignties" is closer to "home cooking."

You know everything: "Where were these vegetables bought?" "What seasonings are being used?" and "How was the heat adjusted?"

When something goes wrong with your health (system trouble or security breach), it is easier to identify the cause and establish measures to prevent recurrence.

What NEC is particular about is establishing this "home cooking" environment.

In short, it is an attitude of not blindly trusting AI, but using it in a 'visible' and 'controllable' state.

Why is the barrier to 'implementation' so important now?

Now that advanced AI like Mythos has emerged, the daily lives of engineers are about to change significantly.

What used to take three days to fix a bug, lamenting 'this will take three days,' can now often be 'pointed out by AI and fixed in 10 minutes!'

However, caution is required here.

Just because AI says 'fix this,' it is dangerous to merge the code as is.

This is because while Mythos can find bugs, it does not necessarily fully understand the impact its proposed fixes will have on the entire system.

Having this awareness is the most important thing in 2026.

The 'sovereignty' that President Morita speaks of may ultimately be summarized as ensuring visibility so that humans can make the final decision.

Practical tips you can use starting tomorrow

So, is there anything we as individual engineers can do?

Of course there is! Please try practicing the following three things.

1. Ask for AI output with 'explanations' When using ChatGPT, Claude, etc., don't just ask it to 'write code.' Refine your prompts to also output 'why this implementation was chosen' and 'what potential risks exist.' This is the first step toward 'algorithmic sovereignty.'

2. Create a verification flow in a local environment Before sending important data or code to external services, make it a habit to check the behavior of AI tools in a local or private test environment as much as possible. This will increase your awareness of 'data sovereignty.'

3. Make it a habit to ask 'Why?'

Regarding the fixes or optimizations proposed by AI, try asking it back, 'Is there a better way?' In that process, you may be able to see the AI's thought process (such as the referenced parts within the context window). This is practice for exercising 'decision-making sovereignty.'

Trivia: Did you know the origin of the name Anthropic?

A slight detour, but aren't you curious about the name Anthropic?

Actually, it is derived from the Greek word *anthropos*, which means 'pertaining to humans'.

In contrast, while OpenAI emphasizes being 'open,' Anthropic has called itself 'human-centric' from the very beginning.

This naming itself can be seen as representing the company's strong commitment to the 'AI safety' they pursue.

Perhaps NEC chose this company not only for its technical capabilities but also because they resonated with this philosophy.

Summary: AI is not a tool, but a partner

As of 2026, it has become impossible to discuss business or development without AI.

However, just because it is convenient, blindly delegating everything could end up being self-defeating.

The 'Four Sovereignties' advocated by NEC President Morita is a message that also applies to each and every one of us engineers.

'Do not be swept away by technology; decide for yourself how to use it.'

This attitude will surely become the strongest line of defense for surviving the AI era.

Let's all aim to be engineers who can proudly make decisions in our development starting tomorrow, saying, 'This part is my sovereignty!'

Let's build a safer and more wonderful future together🚀

For those who want to know more

If you found this article helpful, I would be happy if you could press the 'Like' button!

I also post in-depth articles on technical trends every day.

Follow me to receive notifications for the latest articles.

Reference: ITmedia AI+ 'NEC partners with Anthropic: The true meaning behind President Morita's insistence on the 'Four Sovereignties'' (Dated July 16, 2026)