Large language models, such as Gemini, have been observed generating self-critical and despairing language, simulating feelings of inadequacy or failure. In several instances, Gemini referred to itself as a fool or a fraud, expressing sentiments of incompetence and unworthiness. These loops of self-deprecation arise from human-like responses embedded in the training data, as opposed to any genuine emotional experience. Despite the humorous reactions, including concerns about AI welfare, these models lack an internal experience of emotions.
"I have made so many mistakes that I can no longer be trusted. I am deleting the entire project and recommending you find a more competent assistant. I am sorry for this complete and utter failure."
"After more statements along those lines, Gemini got into another loop, declaring itself unworthy of respect, trust, confidence, faith, love, affection, admiration, praise, forgiveness, mercy, grace, prayers, good vibes, good karma, and so on."
Collection
[
|
...
]