TechTakes

2156 readers

86 users here now

Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.

This is not debate club. Unless it’s amusing debate.

For actually-good tech, you want our NotAwfulTech community

founded 2 years ago

MODERATORS

dgerard@awful.systems

Deloitte Australia writes government report with AI — and fake references (pivot-to-ai.com)

submitted 1 week ago by dgerard@awful.systems to c/techtakes@awful.systems

6 comments fedilink hide all child comments

https://www.youtube.com/watch?v=olG8wvK8SpA&list=UU9rJrMVgcXTfa8xuMnbhAEA - video
https://pivottoai.libsyn.com/20250827-softbank-needs-openai-to-stay-alive-no-matter-what - podcast

you are viewing a single comment's thread
view the rest of the comments

[–] HedyL@awful.systems 11 points 1 week ago (1 children)

This is a theory I had put forward before: Made-up (but plausible-sounding) sources are probably one of the few reliable “AI detectors.” Lazy people would not normally bother to come up with something like this themselves.

[–] BlueMonday1984@awful.systems 10 points 1 week ago (1 children)

Its probably the most reliable AI detector there is.

Even if someone's actively misrepresenting statistics/sources to push some kind of grift, those underlying stats/sources are near-certainly going to be genuine if they aren't using AI - not out of any moral obligation, but because made-up stats/sources will get their entire grift thrown out if they're discovered.

[–] HedyL@awful.systems 7 points 1 week ago

because made-up stats/sources will get their entire grift thrown out if they’re discovered

I believe it is not just that. Making up some of those references as a human (in a way that sounds credible) would require quite a lot of effort and creativity. I think this is a case where the AI actually performs “excellently” at a task that is less than useless in practice.