2026年8月12日
Researchers allege American AI models may be facilitating censorship
要約
Date Reported: 2026年8月12日
場所: グローバル
企業
Anthropic関連
Total individuals affected: Number unknown
課題
検閲 , 表現の自由の否定 , 情報へのアクセス回答
見解が求められた : Yes, by Wall Street Journal
External link to response: (Find out more)
取られた措置: According to the Wall Street Journal: "Anthropic said it worked rigorously to ensure Claude responds in a balanced way. It said its latest models have made significant progress in reducing over-refusals."
情報源のタイプ: News outlet
要約
Date Reported: 2026年8月12日
場所: グローバル
企業
OpenAI関連
Total individuals affected: Number unknown
課題
検閲 , 表現の自由の否定 , 情報へのアクセス回答
見解が求められた : Yes, by Wall Street Journal
External link to response: (Find out more)
取られた措置: According to the Wall Street Journal: "ChatGPT developer OpenAI pointed to its published approach, which says its models default to an objective point of view and 'should never avoid addressing a topic solely because it is sensitive or controversial.'"
情報源のタイプ: News outlet
要約
Date Reported: 2026年8月12日
場所: グローバル
企業
Meta (formerly Facebook)関連
Total individuals affected: Number unknown
課題
検閲 , 表現の自由の否定 , 情報へのアクセス回答
見解が求められた : Yes, by Wall Street Journal
External link to response: (Find out more)
取られた措置: According to the Wall Street Journal, Meta declined to comment.
情報源のタイプ: News outlet
要約
Date Reported: 2026年8月12日
場所: グローバル
企業
Google (part of Alphabet)関連
Total individuals affected: Number unknown
課題
表現の自由の否定 , 検閲 , 情報へのアクセス回答
見解が求められた : Yes, by Wall Street Journal
External link to response: (Find out more)
取られた措置: According to the Wall Street Journal, Google didn’t respond to requests for comment.
情報源のタイプ: News outlet
"Chinese Censorship Is Leaking Into Answers From American AI" 12 August 2026
Ask a chatbot in China about Chinese politics, and it will either clam up or echo the official Communist Party line. New studies show a surprising twist: American artificial-intelligence models are unwittingly doing the same—and researchers say AI companies haven’t done much to fix it...
...American AI companies say they take steps to ensure neutral responses that protect free speech. Still, the oversight board last month published research that found models from OpenAI, Anthropic, Google and Meta were significantly less likely to criticize repressive governments than governments in freer countries...
...Anthropic said it worked rigorously to ensure Claude responds in a balanced way. It said its latest models have made significant progress in reducing over-refusals. ChatGPT developer OpenAI pointed to its published approach, which says its models default to an objective point of view and “should never avoid addressing a topic solely because it is sensitive or controversial.”
Meta declined to comment, and Google didn’t respond to requests for comment...