this post was submitted on 19 Sep 2026
463 points (98.1% liked)

Technology

88163 readers
3385 users here now

This is a most excellent place for technology news and articles.

Our Rules

  1. Follow the lemmy.world rules.
  2. Only tech related news or articles.
  3. Be excellent to each other!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
  9. Check for duplicates before posting, duplicates may be removed
  10. Accounts 7 days and younger will have their posts automatically removed.

Approved Bots

founded 3 years ago
MODERATORS
L3s@lemmy.world
enu@lemmy.world
technopagan@lemmy.world
L4s@lemmy.world
L3s@hackingne.ws
463
AI hallucination of Chinese nuclear components almost led to US military attack (arstechnica.com)
submitted 2 days ago* (last edited 2 days ago) by inari@piefed.zip to c/technology@lemmy.world
65 comments fedilink hide all child comments
you are viewing a single comment's thread
view the rest of the comments
[–] brsrklf@jlai.lu 10 points 2 days ago

You know the thing where AI companies describe their alignment direction as "Honest, Harmless, Helpful"?

The only one that's marketable to the idiots buying AI services is Helpful. So that's always top priority even when it goes against the other two.

Average decider wouldn't stand an "assistant" that sometimes answers "I don't know". So instead agent providers train their already hallucination-prone model to be 100% confident in whatever it says. Who cares if it's true if it looks like it is?