#model-inconsistency

[ follow ]
#climate-risk-modeling #flood-risk #wildfire-risk #ai-content-moderation #hate-speech-detection
Real estate
fromwww.mercurynews.com
4 days ago

Zillow deleting home climate risk scores reveals limits of flood, fire data

Climate-risk property ratings vary widely among models, producing inconsistent flood and wildfire vulnerability assessments that can misprice assets and confuse buyers.
Artificial intelligence
fromFast Company
3 months ago

Researchers studied how Google, OpenAI, Anthropic, and DeepSeek identify hate speech. Here's how they vary

Leading AI moderation models vary widely in identifying hate speech, producing inconsistent, unpredictable, and potentially unfair moderation outcomes.
[ Load more ]