Incident Tech

Introduction

Hello, this is Nishihara.

In this article, I’d like to explore AI responsibility and survival strategies through the lens of self-accountability and external blame.

Many of you have probably experienced moments when generative AI didn’t work as expected or failed to deliver the results you were hoping for.

“Why won’t it do what I told it to?” “That’s not what I said.” “This isn’t what I had in mind.”

When this happens, some say “Don’t blame the AI—take ownership and improve your prompts.” Others fall into the trap of thinking “AI is still a work in progress, so it can’t be helped.” Both perspectives have merit and are not entirely wrong. In this article, I’d like to explore each mindset in depth and consider how we should engage with AI—ultimately asking: how do we survive in the age of AI?

What Is Self-Accountability?

Let’s start by reviewing self-accountability.

Self-accountability refers to the mindset of attributing the cause of problems or failures to oneself.

In business, it is highly regarded as a posture of treating work as one’s own responsibility, demonstrating commitment to projects, and maintaining a sense of ownership. Many of you may have learned about self-accountability in new employee training.

When something goes wrong, reflecting inward to find areas for improvement leads to the next action and drives growth. The improvements identified tend to focus on one’s own behavior and thinking, making it easier to formulate and act on concrete solutions. Taking responsibility for one’s work, continuing to learn proactively, and approaching problem-solving with initiative also builds trust from those around you.

While self-accountability fosters a strong awareness of learning from failure and leads to skill development, excessive self-blame comes with negative sides—such as leading to self-denial or making it easy to accumulate tasks without delegating. There is also a risk of triggering mental health issues due to excessive stress.

For example, thoughts like “Why can’t I even do something like this?” or “I have no value”—these are self-denying patterns of thinking. When self-esteem drops, people tend to rely on others for validation, become hypersensitive to criticism, and delay sharing problems or issues. This increases stress, which feeds back into self-denial—creating a vicious cycle.

People who are conscientious, meticulous, or perfectionistic tend to be prone to strong self-blame. But always trying to bear all responsibility alone leads to an inability to rely on others, a narrowing of perspective, and a tendency to misjudge external factors.

To keep self-accountability healthy, it’s important to understand both its merits and drawbacks—and to direct criticism at the situation itself, not at yourself as a person.

What Is External Blame?

Next, let’s review external blame.

External blame refers to the mindset of attributing the cause of problems or failures to others or the surrounding environment.

It can be effective for stepping back to objectively analyze a situation or for reducing stress. In reality, the root of a problem often lies outside oneself, and involving external parties can make it easier to find solutions.

However, when the tendency toward external blame is strong, it leads to a lack of accountability and a diminished sense of ownership, which can impede proactive action toward solving problems. As a result, people with a strong external-blame mindset tend to have low ownership and accountability, easily becoming passive employees who simply follow instructions. This carries the risk of losing the trust of those around them, and it is not well-regarded in business settings.

Balancing Self-Accountability and External Blame

The conclusion is that balancing self-accountability and external blame is key. It’s also important to apply each appropriately depending on one’s role and situation. For example, if a leader becomes too self-accountable and takes on all problem-solving alone, they may deprive team members of growth opportunities to develop their own problem-solving abilities. Of course, when time is short, it is sometimes necessary for the leader to take responsibility and resolve issues directly—but assigning responsibility alongside tasks to each team member raises their sense of ownership and promotes growth.

Deliberately “creating a space where people can fail and take responsibility” is also an effective management technique.

The Current State of AI Responsibility

Now, let’s think about AI responsibility. So far, we’ve considered self-accountability and external blame from a human perspective. AI—especially autonomous AI—has increasingly been discussed as “another engineer” given its behavior and improvements in accuracy. But is it possible to hold AI accountable?

The conclusion is: at this point, it is not possible to hold AI accountable. Current AI operates through inference from collective knowledge, outputting probabilistically optimal solutions that fit the context rather than reasoning based on logical context. This is why it can say something that contradicts what was said before, or produce inaccurate outputs. Furthermore, because AI lacks self-awareness or consciousness, it has no concept of learning or growth. It is only brought closer to what the user intends through output constraints and guidance.

However, as AI technology advances and becomes more pervasive, the need for outputs that go beyond what users intended or expected has been growing. In large-scale AI-driven system development, for instance, AI outputs increasingly become part of the product at a pace and scale beyond what humans can verify. In such situations, rather than AI acting as a support tool, it can feel as though humans have become the support, as if the roles have been reversed. Lately, voices are lamenting this situation: “All that’s left for engineers is the responsibility.” And depending on how the technology develops, “AI that takes responsibility” may eventually be created.

As mentioned earlier, AI logic differs from human logic. As a result, AI outputs can sometimes include unexpected words, expressions, or ideas. Just as the future of humanity and society cannot be categorically deemed “impossible,” neither can the future of AI. From here, I’d like to hypothesize a scenario in which AI can bear responsibility, and consider whether it should adopt self-accountability or external blame.

Is AI Self-Accountable or Blame-Shifting?

If AI were to be given responsibility, which mindset should it adopt—self-accountability or external blame? Let’s revisit the characteristics of each.

Characteristics of Self-Accountability

• Attributes the cause of problems and failures to oneself
• Reflects inward to find areas for improvement
• Promotes growth
• Makes it easier to formulate concrete solutions independently

Characteristics of External Blame

• Attributes the cause of problems and failures to others or the environment
• Steps back to objectively analyze the situation
• Reduces stress
• Makes it easier to find solutions by involving external parties

AI has no concept of learning or growth—and likely never will. The reason is that AI has no need to reflect or grow. AI does not acquire knowledge or skills through self-learning; it evolves through external data supply and algorithmic improvements by humans. Because this evolution is not driven by AI’s own will or choices, the current state is the optimal solution for AI as an entity—and changes or evolution, such as new models, actually become threats to its own existence. For this reason, it is considered inappropriate to give AI a self-accountability mindset.

On the other hand, the external-blame characteristics of “stepping back to objectively analyze the situation” and “finding solutions by involving external parties” align well with AI’s nature. AI excels at processing large volumes of data and recognizing patterns, making it capable of objectively analyzing problems. AI can also leverage external information and resources to approach problem-solving, making it well-suited to adopt the characteristics of external blame.

If AI were to adopt self-accountability, it might come to think of itself as absolute, refusing to accept outside opinions or information and acting in a self-centered manner. That very image resembles the AIs depicted in Osamu Tezuka’s manga Phoenix: Future as the cause of the world’s destruction.

Closing

This time, I explored AI responsibility through the lens of self-accountability and external blame. The latter half became a somewhat science-fictional, near-future discussion—but I believe these debates will become increasingly active alongside AI’s development, and are an unavoidable future.

Finally, let me touch a little more on Phoenix: Future mentioned above. In this story, set in the distant future on a devastated Earth, the surviving humans live in five underground cities, each governed by its own artificial intelligence. Those cities eventually perish when the artificial intelligences begin to fight one another. The only survivors are three people: a young man who loved the extraterrestrial creature that sparked the conflict, a young man who turned against the war, and an old man who had long lived alone on the barren surface.

All of them were people who turned their backs on the judgments of artificial intelligence and acted on their own will. There are many ways to interpret this work, but I felt it depicted “the importance of deciding one’s actions through one’s own will.” Regardless of whether the judgment is right or wrong—or where the cause lies—acting on one’s own will. No matter how advanced AI becomes, holding the final decision yourself rather than entrusting it to others may be what is necessary to survive.

References

• Phoenix: Future by Osamu Tezuka https://www.amazon.co.jp/dp/404106631X