Post #3046685 · Elektrine

2026-04-15 10:09 UTC

@elduvelle@neuromatch.social @AggroBoy@mastodon.social @codinghorror@infosec.exchange I hear you, and I think you’re right at one level. I think the counter-argument here is: this is obvious to us, but clearly not obvious to many, many users. We understand how this works, and can frame its response, and restrict our usage. Many users do not, and will not, and want to use this stuff, and instinctively want to feel like they are talking to a person of some kind. So there should probably be better guardrails against this kind of nonsense dangerously misleading folk.

Replies (1)

@elduvelle@neuromatch.social 2026-04-15 10:24

@Master_Squinter@infosec.exchange Indeed, this is exactly the problem with the large-scale manipulation performed by the companies that make these products - they try their best to make it look like your are "conversing" with "an individual" instead of sending instructions to a program. and our brains are just used to project intent, will, individuality towards anything that uses language so it has no guardrails against interpreting the output of the program as if it was coming from a person. I personally think the best thing to do is to not use genAI. But of course the companies push it towards users as hard as they can so it must be very difficult to resist for uninformed users. I guess the solution is to try and inform them somehow.. or ban genAI at a governmental level (lol). We definitely can't count on the companies that profit from people thinking that genAI is useful and reliable to tell them that what it outputs is actually not useful or reliable.. #genAI @AggroBoy@mastodon.social @codinghorror@infosec.exchange

Open ##3046686