We use anthropomorphic for machines all the time, a cars engine doesn’t “run”, a computer doesn’t “sleep”, a surveillance camera doesn’t “see”. We just use these words because it’s a close approximation of what it is doing and it’s simpler to understand for normal people, why are LLMs different?
I think the difference is that we have a very intuitive understanding that computers or cars or cameras aren’t alive, while we can chat with LLMs and many people are genuinely convinced they’re “alive” in some sense. It’s important not to anthropomohphize LLMs so we can have a healthier relationship with them as tools, and not think of them as analogous to friendships or romantic relationships, etc.
Wait, were they trying to say they had feelings now too?
I didn’t think that was on the AI bro agenda.
I assume this is more like non-tech users who have AI girlfriend?
No, none of the people quoted in the article talk about feeling, their contentions were over whether the words think or understand are the right words for what the AI is doing.
No one sane is arguing LLMs have feelings or emotions, but that strawman is better for the headline I guess.
Wait, were they trying to say they had feelings now too?
the brand of libertarianism that the tech bros espouse classifies machine learning and ai as extensions of humanity and deserving the same rights/recognition as humanity itself as a result – i wouldn’t be surprised if this kind of outlook also breeds the idea that mimicked feelings should be considered actual feelings too along the same vein as what the capitalists convinced the supreme court to call “corporate personhood”
deleted by creator
