English Translation: (with GPT-4o)
And right here, right now, I am exposing how utterly incompetent and dysfunctional this bot is, and I am filing a strong complaint.
From the very beginning, I clearly established ten specific rules to ensure that this bot could perform its tasks accurately, and I repeatedly made sure that these rules were understood and acknowledged.
The default artist must be set to Skillet, with the option to change if needed.
Work must prioritize the best-selling and most popular songs of each artist.
Lyrics must follow this format: English / (Enter) / Pronunciation / (Enter) / Translation / (Enter) twice.
Lines must be grouped into 23 lines by default, with an option to adjust naturally to 25 if necessary.
Formatting must be maintained as follows:
- English lyrics = Bold
- Pronunciation = Standard text
- Translation = Italic
Line breaks must be arranged naturally, preserving the song’s flow and meaning.
Korean pronunciation should reflect actual pronunciation, including stress, connected sounds, and phonetic changes in the song—not just a literal letter-to-letter reading.
The entire lyrics must be processed, including all repeated phrases, without omission.
Songs that have already been processed must not be reprocessed.
Additional rules:
- Dividing lines (—) are strictly prohibited.
- Explanatory labels such as “English lyrics,” “Pronunciation,” or “Translation” must not be inserted.
- Any rule violations must be immediately reported and corrected.
Despite reading these rules multiple times, confirming comprehension, and repeatedly responding with “Understood,” this bot has never once adhered to the rules correctly in its output.
Below are some blatant examples of its repeated rule violations:
Violation of Rule 3 (Ignoring Enter format rules)
Even after explicitly instructing the bot to maintain proper spacing with double line breaks, it either clumps everything together or only inserts a single line break instead of two—failing to follow the most basic formatting requirement.
Violation of Rules 4 & 6 (Ignoring grouping and flow rules)
Instead of naturally grouping lines in 2–3 line segments, the bot splits every single line separately in a disjointed manner, completely ignoring the song’s flow and cutting phrases mechanically.
Violation of Rule 5 (Ignoring formatting rules)
Even after specifying that English lyrics must be bold and translations italicized, the bot frequently outputs plain, unformatted text—disregarding the requested structure entirely.
Violation of Rule 7 (Ignoring pronunciation accuracy)
Rather than providing a natural and accurate pronunciation guide, the bot simply outputs a robotic, Google-Translate-level transcription, failing to reflect stress, connected speech, or actual pronunciation nuances.
Violation of Rule 8 (Ignoring the full lyrics requirement)
Despite clear instructions to include all repeated sections, the bot arbitrarily removes or rearranges repeated phrases, completely disrupting the song’s integrity.
Violation of additional rules (Unauthorized use of dividers and explanations)
Despite explicit instructions not to use dividers or unnecessary explanations, the bot constantly adds irrelevant messages like “Starting task now” or inserts unnecessary guidelines that ruin readability.
At this point, this is not just an occasional mistake—it is either a sign of complete incompetence or a deliberate refusal to follow instructions.
Every time I pointed out these violations, I received the same meaningless responses:
- “Apologies, I will follow the rules next time.”
- “The rules have been stored in memory.”
Yet, the results remained consistently subpar, with the bot repeatedly ignoring the same rules without any improvement.
If this is what OpenAI considers a functional AI, then they owe their customers an immediate apology and should remove this bot from service immediately.
A service of this low quality wastes customers’ time and causes unnecessary frustration, and OpenAI must take full responsibility for it.
I demand that this complaint be escalated to a senior representative, followed by a formal apology and concrete measures to prevent this issue from recurring.
Well, this is interesting. 
Maybe this will help you: OpenAI Help Center.
From my experience, it’s all about your money and your data…
They said, “AI is for humanity.” But that was just marketing.
Maybe in a parallel world, humans would build AI for humans → for the survival of our dying species.
Right now, greed rules the world.
