"Make it look good, and do it ASAP"—Will AI put an end to vague Japanese in the workplace?

Introduction: Have you ever been troubled by that one phrase?

"Make it look good, and do it ASAP."

I imagine you have been told this at least once. Or perhaps, you have said it yourself.

The person being told this is in a bind. What does "look good" mean? Should it be a rough bulleted list, or a formal document? What does "ASAP" mean—by the end of today, or by the end of the week? If you ask for clarification, you fear the atmosphere will suggest, "Can't you just figure that out?" so you don't ask. In the end, you make a guess, only to be scolded later with, "That's not what I meant."

Now, what happens when generative AI enters this scenario?

I see this change as a small site where the Japanese language is beginning to change slowly. And while this is a story about making the workplace easier to work in, it is also a slightly troublesome one.

Why did work get done with that one phrase?

First, let's consider why "Make it look good, and do it ASAP" was effective.

This instruction is empty in terms of information. Yet, work was still getting done because there was support outside of the words themselves. The boss and the subordinate were in the same workplace, attended the same meetings, and generally knew who the project was for and what was needed by when. As for "looking good," you generally understand the expected tone from your history of working with that boss.

This is not a matter of the Japanese language being vague. Rather, it is a rational energy-saving measure where you don't have to restate the circumstances to someone who already shares them. If you have worked together in the same office for a long time, "Take care of that thing" really does get the message across.

And this omission has another function: the function of entrusting. "Make it look good" is also an expression of intent to leave the fine details of judgment to the other party. There are certainly situations where it is easier to work when you are trusted to handle things rather than having everything specified.

AI is not troubled by this instruction

However, when you throw this instruction at an AI, the situation changes.

The AI does not know your workplace. It wasn't at yesterday's meeting, and it doesn't know who this document is for. Therefore, just saying "Make it look good" can result in nonsensical output.

That is where the human side changes. When asking an AI, you naturally start writing things like: "This is a document for next week's board meeting. One A4 page, up to three main points, and be sure to include the cost." You state the premises, list the conditions, and specify the format.

From here, two paths diverge.

One is a path where this explicitness extends to the boss themselves. The boss acquires the habit of verbalizing conditions in order to give instructions to the AI, and that habit spills over into instructions given to subordinates. They stop saying "Make it look good." Subordinates are saved, and rework is reduced. This is not a bad thing.

The other is a path where the ambiguity is absorbed by the AI. The boss continues to say "Make it look good." The troubled subordinate throws that phrase directly at the AI. Since the AI is not troubled, it returns something that looks plausible. The subordinate tidies it up and submits it, and the boss accepts it, thinking, "Well, I guess this is fine."

At first glance, the latter seems more peaceful. No one is troubled. However, I am much more concerned about this one.

There was a function to the things that troubled subordinates.

Why does this bother me? This is the core of this article.

It is not necessarily the case that a boss who says "make it look good" has a clear image in mind and is just omitting it because it is a hassle to explain. Rather, in many cases, the boss themselves may not have decided what they want yet. "Make it look good" is also a phrase used to pass on something undecided while it is still undecided.

If that is the case, then the subordinate's trouble and their follow-up question, "What kind of image do you have in mind for 'looking good'?" had great significance. That follow-up question forced the boss to verbalize their own thoughts. Only after being asked do they realize, "Oh, I see, I'm using it for the board meeting, so one page is fine." The subordinate's confusion acted as a sensor that detected the boss's lack of thought and sent it back for revision.

AI does not have this sensor. If you throw a vague instruction at it, it will return a plausible-looking deliverable. As a result, the fact that the boss had decided nothing remains hidden until the end. The ambiguity does not disappear; it is processed without anyone noticing it.

It is usually a good thing when "there are no more people in trouble." However, if the people who were in trouble were actually playing the role of finding the parts of the organization that were not thinking, then those people no longer being in trouble also means that role is lost. I think this is worth watching carefully.

Contracts are sometimes written vaguely on purpose.

I would like to write about another realization I have as a lawyer regarding ambiguity.

Contracts are often thought of as documents that have written down everything. In reality, that is not the case. There are many clauses in contracts that are intentionally left vague.

For example, the sentence, "Matters not stipulated in this contract shall be decided upon through good-faith consultation between the parties." This is at the end of almost every contract. This decides nothing. "Notify without delay" also does not write within how many days. "Manage with the care of a good manager" also does not write specifically what should be done.

Why not write everything down? There are several reasons.

First, because you cannot write down everything that will happen in the future. In a transaction that lasts for many years, you cannot list everything that will happen from now on. It is more practical to leave a receptacle called "consultation" for when unforeseen situations occur.

Second, because if you try to write everything down, the contract may not be concluded. If you try to work out every detail, negotiations will drag on. As a result of sticking to points that cannot be settled, the transaction itself may fall through, or costs may become enormous. There are real situations where leaving undecided points intentionally vague makes today's agreement possible.

In other words, ambiguity in contract practice is often not cutting corners, but a technique. It is a tool to move forward without deciding what cannot be decided.

Admittedly, it is not all good. Clauses left vague will always become points of contention when relationships deteriorate. The sentence "consult in good faith" does not prevent disputes. In fact, my job is more often dealing with troubles caused by that very ambiguity. Determining where to write clearly and where to leave it to others is the crux of practical work, and this depends on the case.

"Make it look good, and do it ASAP" also has the same structure. There is bad ambiguity, and there is meaningful ambiguity. I think the problem is not the ambiguity itself, but whether it is being used as a technique or as a cover for not thinking.

How will Japanese in the workplace change?

First, I think the language used to request work will move in an explicit direction. If the habit of giving instructions to AI spreads, the resistance to writing down conditions will decrease. Replacing "ASAP" with "by noon the day after tomorrow" is probably a good change.

On the other hand, language that adjusts relationships will remain. When refusing, when scolding, when asking for something difficult. Both "That's a bit difficult" and "I will consider it positively" are not words for conveying information, but words for not breaking relationships. This function cannot be replaced by explicitization.

If "Make it look good, and do it ASAP" disappears from the workplace, it will not be because AI erased it. AI does not work in the workplace, nor is it chased by deadlines, nor does it worry about its boss. The ones who erased it are we ourselves, who learned to put our thoughts into words in order to give instructions to AI.

I believe it is always the people who live using a language that change it.

Conclusion

The world where people struggle with phrases like "make it look good" and "do it ASAP" may indeed be coming to an end, because AI can now handle vague instructions.

However, there are two ways this could end. Will the ambiguity be verbalized and disappear, or will it be absorbed by the AI and preserved without anyone noticing? If it is the former, the workplace will improve, but if it is the latter, a system will be created where those who haven't made a decision can get away with not making one.

Therefore, if you are told to "make it look good" and "do it ASAP," I think it is worth asking for clarification just once before feeding it into an AI. This is not a tedious verification task, but a job to get the other person to think.

And if you are the one saying "make it look good," you might want to try giving the same instruction to an AI. If you look at what comes back and think, "That's not what I meant," that is the moment you will finally understand a little bit about what you actually wanted. Perhaps it wasn't just your subordinate who was struggling.