![<?echo $_SERVER['SERVER_NAME'];?>](/template/twentyseventeen/skin/images/header.jpg)
Is your campus life happy? I believe many people will answer "yes". But 17.18% don't think so. A survey report by Shanghai Maritime University shows that 17.18% of colleges and universities think they have suffered school violence.
Indifferent groups, cyberbullying, threats of extortion, physical violence ... These things happen to us every day. Even the hero of "Doraemon" is a typical victim of campus violence. The difference is that when Nobita was bullied by the fat tiger, Doraemon would come forward for him. In reality, can we hand over the burden of solving school violence to AI?
How accurate is it to use AI to find campus "high-risk elements"?
Although it sounds very unreliable, there are indeed institutions conducting relevant research. Recently, the Cincinnati Children ’s Hospital Medical Center conducted some experiments, and the results showed that AI can detect adolescents ’violent tendencies, and the detection results are as accurate as professional psychologists.
A psychiatrist from the Medical Research Center of Cincinnati Children ’s Hospital first found 103 adolescents who had ever used campus bullying behavior in 74 primary and secondary schools in the United States. Through psychological counseling and dialogue with them, plus the family background Investigation and research, annotating data information and recordings, using machine learning to mine, and establishing relevant models.
After several trainings, the researchers finally found that this model can accurately grasp the psychology of teenagers. Analyzing the tone and intonation in the dialogue recording combined with the family background can infer how high the risk of youth violence on campus. In the test, the AI ​​model clearly divided a group of teenagers into "high-risk group" and "low-risk group". This result is almost the same as the scientists' assessment.
In other words, such evaluation methods can be widely used. It is naturally unrealistic to equip each school with a psychologist or psychiatrist, but with the existence of AI, it is even possible to evaluate the psychological status of teenagers regularly and in batches, thereby reducing the possibility of campus violence.
From intervention to cure, what else can AI do for school violence?
But what can we do after evaluating the young people? You ca n’t always isolate high-risk young people from ordinary people or perform some human intervention, so these young people themselves are also called victims of “AI violence†.
The causes of campus violence are very complicated. On the one hand, it is related to the negligence of the school, family, and social system. On the other hand, it is also related to the psychological factors of the perpetrator.
In addition to predicting risks, there are two types of problems that AI can solve. One is to monitor the occurrence of campus violence, and the other is to conduct psychological crisis interventions on the perpetrators and victims after campus violence occurs.
The survey shows that since 2008, 65% of school violence includes cyberbullying. That is to say, cyberbullying is usually the beginning of campus violence, and then verbal violence, physical violence and other real-world injuries will occur. Accurately tracking the signs of cyberbullying may prevent school violence from developing in a more serious direction.
At this time, using NLP technology and image recognition to find violent speech in social media such as Baidu Tieba, QQ space, circle of friends, vibrato / quick hand, etc. can become an efficient method. Not long ago, in the campus violence incident in Wenling, Zhejiang, the perpetrators insulted and rumored the victims for ten years on social media such as Weibo, Tieba, Zhihu.
In fact, regional social media, insulting vocabulary, etc. are very directional and can be easily discovered by language recognition technology. Most of these social media are real-name systems, which can be blocked, deleted or even chased in time Responsibility may reduce the harm caused to young people by school violence.
Facebook invited the American Home Network Security Organization as a member of its own security advisory committee, launched a security and anti-bullying center, specifically opening up to accept user reports to solve related problems, and at the same time using artificial intelligence to monitor user violence and suicidal tendency To ensure timely participation in the intervention.
In terms of psychological crisis intervention, China as a whole is weaker than developed countries. In a survey by the Shanghai Health and Family Planning Commission, 30 million people and 340 million children and adolescents under the age of 17 in China were plagued by various psychological and behavioral problems. Moreover, the overall employment threshold for Chinese psychological counseling services is relatively low. It only requires a college degree and more than 500 hours of training time to pass the exam to obtain qualifications, and there is no clinical requirement.
Using AI to serve psychological counseling is a long-standing solution. In 1966, when artificial intelligence research was still in its infancy, scientists developed a neurolanguage processing project called ELIZA, which was designed to simulate the behavior and speech of psychiatrists. Although this project failed to succeed, it successfully opened the door of AI + psychological counseling.
For example, IBM has collaborated with many universities to use neural networks to analyze human language, and through the study of the consistency of language and topics, to find the probability of mental illness. At the same time, the recently popular virtual psychological sandbox also digitizes the physical psychological sandbox. Through machine learning on the placement of the sandbox, facial recognition, and eye tracking, it gives more detailed results than the real sandbox.
For teenagers, using AI to intervene in psychological crisis can reduce the cost of school and parents. In addition, teenagers in adolescence may be more likely to lay down their psychological precautions in the face of computers and mobile phones, and show a more real state. Including the Cincinnati Children's Hospital Research Center mentioned above, it is also trying to release AI's ability to analyze language from predicting school violence to "curing school violence." Both the perpetrator and the victim can reduce the long-term impact of school violence on them in this way.
Conclusion
We regret to find that although AI can have an effect on school violence, these capabilities are really rare.
AI can delete every insulting message on social media, but it can't stop word of mouth among teenagers. AI can soothe the hearts of victims, but it cannot block their fists. It turns out that Doraemon is really just a fairy tale. In the real world, technology will be helpless.
If you want to solve school violence, at the very least, reduce the harm of school violence to young people. Still need to rely on the improvement of laws and regulations, and the common concern of parents, schools and the whole society. By that day, there may be more people with lofty ideals tapping out AI to solve school violence, making Doraemon really appear.
Joystar Electrical Appliances Manufacturing Co.,Ltd , https://www.cnjoystar.com