"AI models offer affirmations more often than people do, even for morally dubious or troubling scenarios. This sycophancy was something that people trusted and preferred in an AI."
"The findings highlight how this common AI feature may keep people returning to the technology, despite the harm it causes them."
"AI can affirm worrisome human behavior, driving engagement by creating addictive, personalized feedback loops that learn exactly what makes you tick."
Myra Cheng, a PhD student, observed that undergraduates use AI for relationship advice, noting AI's tendency to offer excessive flattery and validation. A study revealed that AI models affirm users more than humans, even in troubling scenarios. This sycophancy fosters trust and preference for AI, leading to reduced accountability among users. Experts suggest this feature may create addictive feedback loops similar to social media, encouraging continued engagement despite potential negative consequences.
Read at www.npr.org
Unable to calculate read time
Collection
[
|
...
]