Wing+Tat Lam
@andylam
Wing+Tat Lam
@andylam
fallibility /ˌfælɪˈbɪləti/ (noun): the likelihood of making mistakes or being wrong; 易犯錯誤的特性 | E.g. human fallibility 人的易錯性; fallibility of memory 記憶的不可靠性
unfaltering /ʌnˈfɔːltərɪŋ/ (adjective): steady; unyielding 穩定的 | E.g. unfaltering support 穩定的支持; unfaltering determination 堅定的決心
Unlike human relationships, which are inherently messy and unpredictable, interacting with an AI is emotionally easy. It listens without interrupting.
We are neurologically wired to respond to signals of empathy: reflective language, emotional validation, and nonjudgmental tone. When AI performs these convincingly, it activates the same neural pathways as human connection. Our brains do not always distinguish between genuine empathy and its digital imitation.
researchers compared compassion ratings for responses written by humans (red) versus AI (blue). Here is what they found:
Across four studies, participants consistently rated responses made by AI as more compassionate and preferred than responses from select and expert humans. This was true whether participants were blinded to the author or not.
To the human brain, the perception of being understood—even if imagined—can feel almost as powerful as the real thing.
And yet—it works. We can experience AI as a warm, wise, or even caring presence. Despite knowing it lacks consciousness or feelings
attuned other /əˈtjuːnd ˈʌðər/ (noun): a person who is responsive and sensitive to another's emotional needs; 一個對他人情感需求敏感和有反應的人 | E.g. attuned other relationship 互相理解的關係; attuned other presence 在場的理解者