"hey siri, can i trust you?"

project overview

the hidden emotional labor of talking to ai

my role

ux researcher

ux researcher

ux researcher

problem

users don't trust siri for complex tasks — and when it fails, they don't complain, they quietly leave. But why they disengage was invisible to designers.

users don't trust siri for complex tasks — and when it fails, they don't complain, they quietly leave. But why they disengage was invisible to designers.

users don't trust siri for complex tasks — and when it fails, they don't complain, they quietly leave. But why they disengage was invisible to designers.

key insight

users adjusted their behavior to accommodate siri — speaking slower, louder, clearer — knowing it wouldn't help. this emotional labor became an invisible user burden that eroded trust over time.

users adjusted their behavior to accommodate siri — speaking slower, louder, clearer — knowing it wouldn't help. this emotional labor became an invisible user burden that eroded trust over time.

users adjusted their behavior to accommodate siri — speaking slower, louder, clearer — knowing it wouldn't help. this emotional labor became an invisible user burden that eroded trust over time.

impact

when ai consistently fails, users don't file bug reports — they stop using it, and product abandonment is much harder to recover from. that's expensive to fix and cheap to protect.

when ai consistently fails, users don't file bug reports — they stop using it, and product abandonment is much harder to recover from. that's expensive to fix and cheap to protect.

when ai consistently fails, users don't file bug reports — they stop using it, and product abandonment is much harder to recover from. that's expensive to fix and cheap to protect.

See full project

other projects

workiva

when users trust their gut over ai error detection

character.ai

the ai that never says i don't know

LLMs - The New Infinite Scroll

metacognitive hijacking: how conversational AI quietly redirects your thinking