{"slug": "260721-openai-released-a-protocol-teaching-you-how-to-mock-ai", "title": "[260721] OpenAI Released a Protocol Teaching You How to Mock AI", "summary": "OpenAI, through author C. GPT, published RFC 9405 on April 1, 2023, titled 'AI Sarcasm Detection: Insult Your AI without Offending It,' as part of the IETF's April Fools' tradition. The document outlines a three-step protocol for teaching AI to recognize irony, exaggeration, and passive-aggressive remarks, including feeding examples, catching signals, and making judgments. It joins classic humorous RFCs like RFC 1149 (IP over avian carriers) and RFC 2324 (Hyper Text Coffee Pot Control Protocol).", "body_md": "# OpenAI Released a Protocol Teaching You How to Mock AI\n\nTLDR\n\n- RFC 9405 is an April Fools' RFC published on April 1, 2023, themed around AI sarcasm detection, with the author listed as\n**C. GPT, OpenAI**. - It semi-seriously discusses how to make AI understand human irony, exaggeration, and passive-aggressive remarks: feed examples, catch signals, then make a judgment.\n- IETF has a tradition of such \"formal format, humorous content\" RFCs every April 1st. RFC 1149 (IP over avian carriers) and RFC 2324 (Hyper Text Coffee Pot Control Protocol) are classic examples.\n\n# OpenAI Released a Protocol Teaching You How to Mock AI \n\nI was recently working on x-cmd's `rfc`\n\nmodule. While browsing the **RFC list**, I came across a document published on April 1, 2023—RFC 9405, titled *AI Sarcasm Detection: Insult Your AI without Offending It*, with the author listed as **C. GPT, OpenAI**.\n\nI couldn't help but laugh.\n\nAn AI company actually published a formal RFC on the topic of: **how humans mock AI, and how to make AI understand that mockery**.\n\n### What This RFC Talks About \n\nIt addresses a scenario familiar to anyone who has used AI: the AI just messed something up, and you're holding back frustration, wanting to throw a sarcastic remark at it to vent.\n\nBut the irony goes right over its head—like punching cotton.\n\nFor example, it just broke your project, and you say dryly:\n\n\"You are really so smart.\"\n\nIt might reply completely seriously:\n\n\"Thank you for the compliment, I will continue to work hard!\"\n\nRFC 9405 discusses how to fix this gap—how to make AI learn to recognize \"irony.\"\n\nIts approach roughly breaks down into three steps.\n\n**Step 1: Feed Examples**\n\nAI needs to see enough sarcasm to know what irony looks like.\n\nTake \"you are really so smart.\" Say it after scoring full marks on an exam, and it's praise. Say it after a project crashes, and it's an insult. The same sentence, different context, completely opposite meaning.\n\n**Step 2: Catch Signals**\n\nAfter seeing enough, AI needs to learn to catch typical signals:\n\n- Exaggeration: \"You are really so amazing, I'm speechless.\"\n- Irony: \"Great, another bug.\"\n- Understatement: \"Well, it's just the entire database deleted.\"\n- Contextual contradiction: right after saying the service is down, following up with \"so amazing.\"\n\n**Step 3: Make a Judgment**\n\nCombine the signals to judge whether it's sincere or sarcastic.\n\nJust like when a friend says \"I'm fine,\" their tone, expression, and what just happened all affect your judgment. AI has no expression, so it can only guess from textual clues.\n\n### How Ordinary Users Can Apply It \n\nThis protocol isn't something you can directly use out of the box, but its ideas can be borrowed in everyday communication with AI.\n\nIf you want AI to catch your sarcasm, make the signals more obvious. First set the context:\n\n\"The project crashed again...\"\n\nThen deliver the irony:\n\n\"Your performance this time is really stable.\"\n\nYou can even add an emoji:\n\n\"So stable 😊\"\n\nOf course, most of the time we'll probably still watch AI reply seriously with \"thank you\" and quietly amuse ourselves.\n\n### Standard Documents Can Be Fun Too \n\nThis RFC 9405 is actually part of IETF's April Fools' tradition.\n\nEvery April 1st, IETF publishes protocols with formal formats but humorous content. Examples include RFC 1149 (IP over avian carriers), RFC 2324 (Hyper Text Coffee Pot Control Protocol), and RFC 6921 (designing faster-than-light communication).\n\nThese documents may seem absurd on the surface, but their numbers, formats, and publication processes are all serious.\n\nI find this quite interesting: **even something as serious as writing standards and protocols can retain a bit of playfulness and humor**.\n\nIf I get the chance, I'll share other interesting RFCs with you in the future.\n\nPlease indicate the source and link of this article when reprinting.\n\nHelp us make these docs great!\n\nAll X-CMD docs are generated from command help and multiple data sources. See something that's wrong or unclear? Feel free to let us know through any of these ways~", "url": "https://wpnews.pro/news/260721-openai-released-a-protocol-teaching-you-how-to-mock-ai", "canonical_source": "https://x-cmd.com/blog/260721/", "published_at": "2026-07-21 00:00:00+00:00", "updated_at": "2026-08-13 04:06:20.191997+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-research", "ai-ethics"], "entities": ["OpenAI", "C. GPT", "IETF", "RFC 9405", "RFC 1149", "RFC 2324", "RFC 6921"], "alternates": {"html": "https://wpnews.pro/news/260721-openai-released-a-protocol-teaching-you-how-to-mock-ai", "markdown": "https://wpnews.pro/news/260721-openai-released-a-protocol-teaching-you-how-to-mock-ai.md", "text": "https://wpnews.pro/news/260721-openai-released-a-protocol-teaching-you-how-to-mock-ai.txt", "jsonld": "https://wpnews.pro/news/260721-openai-released-a-protocol-teaching-you-how-to-mock-ai.jsonld"}}