0 0
Read Time:3 Minute, 21 Second

Electrical engineering team combine the power of humans and AI to create a prototype device that sounds an alarm when an indecent image appearsBut an observer says there is no law to regulate the use of such objects, which could be used by censors, or to protect the data they collect

A device that can detect when a man is watching pornography by “reading his mind” has been developed in China, according to the research team in Beijing behind the project.

The device, which could speed up the work of censors trying to spot indecent images on the Chinese internet, is worn on the head by the subject and can pick up a spike in the brainwaves triggered by explicit content, according to the researchers.

Fifteen male university students aged between 20 and 25 volunteered to wear the item while in front of a computer screen.

Each time a sensitive photo appeared, an alarm went off.

The prototype device proved that human-machine collaboration was feasible “for bad information detection”, said Xu Jianjun, director of the electrical engineering experiment centre at Beijing Jiaotong University, in a peer-reviewed paper published in the domestic Journal of Electronic Measurement and Instrumentation on June 13.

In some countries, including China, watching porn is a crime. Chinese authorities and companies have used artificial intelligence to flag dubious content.

They also employ a large number of professional censors known as jian huang shi, or “porn appraiser”, to monitor videos and photos posted on the internet or social media platforms.
However, AI sometimes makes mistakes because of the limits of image recognition algorithms. And humans grow tired quickly.

The new device is an attempt to use the combined power of human and machine to achieve greater precision and efficiency, according to Xu and his colleagues.

Despite the rapid development of AI in recent years, human eyes and brains still outperform machines when detecting pornographic content, especially for images containing a complex background, the researchers said.

A nude photo hidden in a stream of normal pictures could raise the eyebrows of an observer, even if it appeared on screen for half a second, they found.

Using the helmet, a jian huang shi only needed to sit in front of a screen, and the device would take over to screen a large number of photos non-stop until the censor blinked.

The developers said the new device could automatically adapt to the brainwaves of a human censor and could filter out noises in the brain signal caused by emotion, weariness or other thoughts.

But there is room for improvement in the technology, according to the researchers.

Although the device detected nearly all the pornographic content, it occasionally raised a false alarm.

To meet legal requirements in China, Xu’s team had to cover up the most sensitive areas of a sample image, which could reduce the visual impact. They said they also struggled to find sufficient training material.

Many jian huang shi are women but no female volunteers took part in this programme, leaving unanswered the question of whether gender affects the device’s performance.

In the paper, the researchers said more training was needed to improve the performance of the device, which recorded overall accuracy over 80 per cent.

A researcher studying brain-machine interface at the University of Science and Technology of China in Hefei, Anhui, said the technology could cause ethical problems, such as privacy infringement.

“There is no law to regulate the use of such devices or protect the data they collected,” said the researcher who asked not to be named.

The People’s Liberation Army has funded the development of a mind-reading helmet for soldiers to communicate with smart weapons, according to openly available information.

Some factories in China used brain surveillance devices to prevent work accidents by monitoring workers’ attention and emotions.

Another recent study by Chinese scientists found that a mind-reading robot could work side-by-side with humans to increase the productivity of an assembly line

About Post Author

Treadstone 71

@Treadstone71LLC Cyber intelligence, counterintelligence, Influence Operations, Cyber Operations, OSINT, Clandestine Cyber HUMINT, cyber intel and OSINT training and analysis, cyber psyops, strategic intelligence, Open-Source Intelligence collection, analytic writing, structured analytic techniques, Target Adversary Research, cyber counterintelligence, strategic intelligence analysis, estimative intelligence, forecasting intelligence, warning intelligence, threat intelligence
Happy
Happy
0 %
Sad
Sad
0 %
Excited
Excited
0 %
Sleepy
Sleepy
0 %
Angry
Angry
0 %
Surprise
Surprise
0 %

By Treadstone 71

@Treadstone71LLC Cyber intelligence, counterintelligence, Influence Operations, Cyber Operations, OSINT, Clandestine Cyber HUMINT, cyber intel and OSINT training and analysis, cyber psyops, strategic intelligence, Open-Source Intelligence collection, analytic writing, structured analytic techniques, Target Adversary Research, cyber counterintelligence, strategic intelligence analysis, estimative intelligence, forecasting intelligence, warning intelligence, threat intelligence