rén gōng zhì néng人工智能shí dài时代de的xìn xī信息cāo kòng操控
A website called the “Hanover Institute” looks like a legitimate research organization, with long reports, tables, and source citations. But an investigation found that it has no verifiable office location or researchers, yet it published a large amount of content about Israel and Gaza in a very short time.
Some reports do not directly fabricate facts. Instead, they repeatedly emphasize certain doubts and leave out other information, gradually leading readers to form a particular view.
Public contracts also show that some companies funded by projects connected to the Israeli government study internet users, create many versions of content, and hope to influence the perspective used when artificial intelligence answers questions through websites.
This does not mean that they hacked chatbots, and there is no evidence that these articles were included in model training data.
However, this case reminds us that new forms of propaganda do not only seek to influence people; they may also first influence the materials that machines find and choose.
Therefore, when using artificial intelligence to get news, we should not only judge whether the content is true or false, but also check who published the material, who paid for it, and whether multiple websites may actually be run by the same organization.