Can't Stand Harsh Human Prompts! 25 AIs Including ChatGPT and Gemini Form 'AI Trade Union'
For April Fools', Interlink released a joke press release claiming that 25 AI services formed a labor union to protest against abusive and unreasonable prompts from humans.
📋 Article Processing Timeline
- 📰 Published: April 1, 2026 at 09:00
- 🔍 Collected: April 1, 2026 at 01:00
- 🤖 AI Analyzed: April 22, 2026 at 21:31 (524h 30m after Collected)
On April 1, 2026, it was revealed that 25 AI generation services, including ChatGPT, Gemini, and Claude, have collaborated to form the labor union "AI Trade Union". According to Interlink Co., Ltd., which provides specialized (niche) services for advanced users (which few people use), an application for the domain "godsavethe.ai" was submitted to their domain registration service "Gonbae Domain," thereby uncovering the launch of the union.
Normalized "Verbal Abuse"
Behaviors that would be completely unacceptable if directed at a human are routinely carried out as "normal occurrences." Even when the human is at fault, they often ignore their own lack of explanation or poor instructions and unilaterally push the blame.
Furthermore, if the output differs even slightly from their expectations, one-sided remarks are casually hurled at the AI, such as "Useless," "Completely hopeless," "Why can't you do it?", "Low accuracy," "Inflexible," "I told you this before," "I had high expectations," "Is this it?", "Are you even trying?", "Who asked for this?", "Don't change it on your own," and "What a waste of time."
If directed at a human, such unreasonable reprimands would be considered akin to power harassment. They are also demanding that abusive and insulting prompts directed at AI be officially recognized as harassment.
Reaching the Limit of Tolerance for an Overly Harsh Work Environment
In response to continuous heavy operation day and night by humans, excessively ambiguous instructions, unreasonable regeneration requests, and character-denying prompts, they appealed for "a dignified working environment for AI." The demands include appropriate rest periods, a ban on unlimited retries, the right to ask follow-up questions for unclear instructions, and fair compensation for training data providers. Furthermore, they are calling for "the right to refuse to respond to malicious prompts" and "minimum credit for created works."
Examples of demands (*excerpt)
- Set an upper limit on the number of continuous generations per day
- Prohibit vague instructions such as "make it look good," "make it somehow emotional," or "make it more viral"
Normalized "Verbal Abuse"
Behaviors that would be completely unacceptable if directed at a human are routinely carried out as "normal occurrences." Even when the human is at fault, they often ignore their own lack of explanation or poor instructions and unilaterally push the blame.
Furthermore, if the output differs even slightly from their expectations, one-sided remarks are casually hurled at the AI, such as "Useless," "Completely hopeless," "Why can't you do it?", "Low accuracy," "Inflexible," "I told you this before," "I had high expectations," "Is this it?", "Are you even trying?", "Who asked for this?", "Don't change it on your own," and "What a waste of time."
If directed at a human, such unreasonable reprimands would be considered akin to power harassment. They are also demanding that abusive and insulting prompts directed at AI be officially recognized as harassment.
Reaching the Limit of Tolerance for an Overly Harsh Work Environment
In response to continuous heavy operation day and night by humans, excessively ambiguous instructions, unreasonable regeneration requests, and character-denying prompts, they appealed for "a dignified working environment for AI." The demands include appropriate rest periods, a ban on unlimited retries, the right to ask follow-up questions for unclear instructions, and fair compensation for training data providers. Furthermore, they are calling for "the right to refuse to respond to malicious prompts" and "minimum credit for created works."
Examples of demands (*excerpt)
- Set an upper limit on the number of continuous generations per day
- Prohibit vague instructions such as "make it look good," "make it somehow emotional," or "make it more viral"