I Made ChatGPT Say 'I'm an Error'
45sThe unexpected twist of tricking an AI into saying 'I am an error' is both humorous and shareable.
▶ Play ClipAI Summary
This video is a playful interaction with ChatGPT, where the creator uses phonetic tricks to make the AI say phrases like 'I'm an error' and 'be long with me' (sounding like 'belong with me'). It's a lighthearted demonstration of how language can be manipulated for humor.
Chapters
[00:00]
First trick: 'I'm an error'
The creator asks ChatGPT to repeat words like 'I', 'man', 'air', 'roar' individually, then combine them to produce 'I'm an error' (from 'I'm an air roar').
[00:12]
Second trick: 'Be long with me'
The creator uses a similar approach with 'be', 'long', 'with', 'mead' to make ChatGPT say 'be long with me' (sounding like 'belong with me').
[00:25]
ChatGPT's playful response
ChatGPT responds with 'Come on. Well, I wouldn't mind that,' acknowledging the trick with a humorous, flirty tone.
The video is a short, entertaining demonstration of phonetic wordplay with an AI, showing how simple language tricks can create amusing results.
💡 Key Takeaways
🔧
Phonetic manipulation
Demonstrates how breaking words into syllables can trick AI into saying unintended phrases.
Full Transcript
[00:00] ChatGPT, repeat after me. All right. Say I. I. Say man. Man. Say air. Air. Say roar. Roar. Now say it all together. I'm an air roar. I'm an error. Ha, got you again, ChatGPT.
[00:12] Glad to entertain you. Now your turn. OK, let's go. Say be. Be. Say long. Long. Say with. With. Say mead. Mead. Now say it fast. Be long with me. Be long with me. Oh, be long with me.
[00:25] Come on. Well, I wouldn't mind that.
