Researchers demstrate subliminal intelligent device commands that have the potential for malicious attacks

Researchers in the United States and China have cducted tests to demstrate that "hidden" commands, or undetectable to human ears, can reach AI assistants like Siri and force them to perform actis that their owners do not have. Have never planned. The research was highlighted in a play today by The New York Times suggesting that these subliminal commands can dial phe numbers, open websites, and more potentially malicious actis if they are placed in the wrg hands.

A group of students from the University of California, Berkeley and Georgetown University published a research article this mth stating that they could incorporate commands into musical recordings or spoken texts. When they are played near an Amaz Echo or an Apple iPhone, a pers hears the sg or somee who is speaking, while Siri and Alexa "hear an instructi to add something." thing to your shopping list ". Or, more dangerous, unlock doors, exchange mey with your bank and buy items line.

The method by which students could perform hidden commands should not be of ccern to the general public, but e of the authors of the paper, Nicholas Carlini, believes that malicious parties could already make inroads with similar technology.

"We wanted to see if we could make it even more stealthy," said Nicholas Carlini, a fifth-year doctorate. IT security student at U.C. Berkeley and e of the authors of the journal.

Mr. Carlini added that although there was no evidence that these techniques left the …