Researchers have developed a new way to prevent AI systems from tracking you, but the solution is as dire as the problem it solves. In general, the system predicts which words will be spoken in the next sentence, producing sounds that do not prevent other people from understanding the words, but leave smart devices designed to pursue capitalism in complete chaos.
Science.org reports that a research team has developed a new way to combat AI voice recognition systems like Amazon Echo devices that keep listening to users. The new technology is called Neural Sound Camouflage, and it can confuse AI assistants by creating special background noises when you speak.
Amazon Echo (AP)
The system uses “enemy attack” to combat AI eavesdropping. The machine learning technique uses it to manipulate the sounds to trick the AI that the sounds are wrong with something else. Often, one AI is used to confuse another. However, the process is complex, and the machine learning AI has to process the entire audio clip before it learns how to tune it, making it difficult to use in real time, like in a speech.
Researchers taught the machine learning system to predict the future. The model is trained in hours of conversation recording, continuously processing the remaining 2 seconds of sound and covering up possible things to say. Science.org cites “enjoy a big feast” as an example where AI can predict the response.
The AI listens to what is being said and makes irritating sounds with a series of sentences likely to follow. To the human ear, sound is like background noise and spoken words are easy to understand but not understood by machines.
Chikyo, a computer scientist at Columbia University who led the study, said this is the first step in tackling AI. “Artificial intelligence collects data about our voices, faces and actions. “We need next-generation technology that respects our privacy.”
“There’s something good about how it combines future prediction, a classic machine learning problem, with another machine learning problem, machine learning,” commented Andrew Owens, a computer scientist at the University of Michigan in Ann Arbor.
Read more here at Science.org.
Source: Breitbart
