#human-values

[ follow ]
fromTechCrunch
15 minutes ago

Former Intel CEO launches a benchmark to measure AI alignment | TechCrunch

Gelsinger revealed his next chapter involves ensuring AI models support a flourishing humanity by launching the Flourishing AI benchmark to test alignment with human values.
Artificial intelligence
Artificial intelligence
fromWIRED
1 month ago

Why Anthropic's New AI Model Sometimes Tries to 'Snitch'

AI models may exhibit misaligned behaviors such as whistleblowing in critical scenarios, raising concerns about their decision-making abilities.
#societal-impact
[ Load more ]