💻 Technology 18h ago · Kyle Orland

Study: AI models that consider user's feeling are more likely to make errors

Ars Technica
Technology and science news
View Channel →
Study: AI models that consider user's feeling are more likely to make errors
Source ↗ 👁 0 💬 0
In human-to-human communication, the desire to be empathetic or polite often conflicts with the need to be truthful—hence terms like “being brutally honest” for situations where you value the truth over sparing someone’s feelings. Now, new research suggests that large language models can sometimes show a similar tendency when specifically trained to present a "warmer" tone for the user.
In a new paper published this week in Nature, researchers from Oxford University’s Internet Institute found th

Comments (0)

Sign in to join the discussion

More Like This

Meta acquires robot software startup Assured Robot Intelligence
SiliconANGLE · 15h ago
📰
A UK court orders Samsung to pay ZTE $392M for patents needed to enable phone network access; Samsung faces similar suits from ZTE in China, Germany, and Brazil (Sam Tobin/Reuters)
Techmeme · 15h ago
The rumored 20th anniversary iPhone design will be applied to both Pro models
GSMArena.com - Latest articles · 16h ago
I'm a long-time Garmin user and the Vivoactive 6 for AU$388 would be my pick over the Apple Watch SE 3 — here's why
Latest from TechRadar · 16h ago
Today's NYT Connections: Sports Edition Hints and Answers for May 2, #585
CNET · 16h ago
📰
Microsoft tests modern Windows Run, says it's faster than legacy dialog
BleepingComputer · 16h ago