welcome
Ars Technica

Ars Technica

Technology

Technology

Over half of LLM-written news summaries have “significant issues”—BBC analysis

Ars Technica
Summary
Nutrition label

86% Informative

BBC analyzed four large language models for news summaries.

ChatGPT-4o, Microsoft Copilot Pro , Google Gemini Standard , and Perplexity performed poorly.

The results found inaccuracies, misquotes, and/or misrepresentations of BBC content in a significant proportion of the tests.

VR Score

91

Informative language

94

Neutral language

61

Article tone

formal

Language

English

Language complexity

67

Offensive language

not offensive

Hate speech

not hateful

Attention-grabbing headline

not detected

Known propaganda techniques

not detected

Time-value

medium-lived

Affiliate links

no affiliate links