After engaging extensively with text-based GAIs, I have arrived at two key AI literacy recommendations.
Recommendation one
Adopt a zero-trust stance toward GAI outputs. Remind yourself and others that GAIs are bullshit machines — systems designed to mimic human speech, not pursue truthfulness. Treat all their answers with skepticism, both in terms of factual accuracy and reasoning.
Recommendation two
Recognize that GAIs are performative systems without agency, empathy, or responsibility. While they simulate understanding and rapport, no real bonds or relationships between you and them will emerge. Be wary of their sycophancy and personalized alignment by cross-checking their reactions to you across different GAIs.
Implication
None of this means GAIs are useless. When understood and treated as alien tools, they can be fun, inspiring, and even practically helpful. But they are not reliable — and they are not friends.
