"What did you have for dinner last night, and why did you end up eating that?"
We asked real people and several AI models the same question. You'll see 10 responses — your job is to figure out which ones came from a human and which were generated by AI.
10
Responses
?
Human or AI
focaldata | quiz
Response 1 of 10
The question we asked
"What did you have for dinner last night, and why did you end up eating that?"
focaldata | quiz
Full Breakdown
What gives AI away?
Here's what to look for when spotting AI-generated responses
🧹
Too clean and structured
AI responses tend to follow a neat pattern: state the meal, explain the reasoning, wrap up tidily. Real people ramble, go on tangents, and contradict themselves.
🏪
No real-world specifics
Humans name brands (M&S), real people (Meera Sodha), and specific places (5-a-side after work). AI keeps things generic — "a bag of arborio rice" rather than telling you where it's from.
⚖️
Suspiciously balanced reasoning
AI loves to sound reasonable: "quick, healthy, and satisfying" or "warm and comforting." Humans are messier — they eat something because it was the only veggie option, or because the pesto was about to go off.
🎭
Missing emotional texture
Real responses carry personality — sarcastic quotes around "creative", exclamation marks about a favourite recipe, or the story of negotiating dinner with a partner. AI responses feel performed rather than lived.
Want deeper insight into AI and human behaviour?
Focaldata fuses AI with human expertise to help you understand what real people think, feel, and do. Talk to our team about how we can help with your research.