Why can't AI technology catch the human trafficker "Mei Yi" who is so developed nowadays?
This is one of the most discussed questions among netizens recently.
During this period of time, a news about the human trafficker "Mei Yi" has been circulating on social media.
"Remember Mei Yi's appearance" and "Find Mei Yi"... accompanied by the latest color portrait of "Mei Yi", the whole internet is searching for "Mei Yi".
However, no one knows who Mei Yi is, she remains a mystery.
All we know is that Mei Yi has been suspected of trafficking 9 children in the past decade, and the happiness of 9 families has been destroyed by her: over a decade has passed, most of the children are still missing, and one of the children's fathers committed suicide by jumping out of the window of a train toilet after three years of unsuccessful search, and the mother subsequently remarried.
It is the wish of the whole nation to bring "Mei Yi" to justice as soon as possible.
However, the color portrait of "Mei Yi" that has gone viral on the internet was quickly debunked by the authorities. The official Weibo account of the Child Missing Information Emergency Release Platform of the Ministry of Public Security stated that the image is not an official release, and the Guangdong police are still actively searching for the whereabouts of the other 7 children.
Lin Yuhui, a portrait expert, said in an interview with the media that the second portrait was drawn based on the description provided by Mei Yi's ex-boyfriend, and it was prematurely released online: "The old man clearly told me that the similarity of the drawing is already very high, close to 90%".
"Mei Yi" is indeed real and still at large, she may be lurking around us. However, it is like finding a needle in a haystack to find her relying on manpower alone, and it is quite difficult to identify her.
"To be honest, this appearance really doesn't have much recognition, it feels like an ordinary middle-aged woman," commented netizens.
According to relevant statistics, there are approximately 70,000 missing children in China each year, and only about 5% of them are found. Although child trafficking cases in our country have shown a "low occurrence and high resolution" trend, the difficulty of "searching for relatives" is still extremely high. So, in this era of advanced technology, is there any technology that can help these families searching for their missing children?
AI has become a new hope for finding trafficked children and suspects.
Can AI recognize "Mei Yi"?
It is almost impossible to find a child in the vast sea of people with human eyes alone.
According to previous media reports, in 2015, the police brought back more than 300 suspected photos through child portraits and organized lost parents to identify them, identifying 176 suspected individuals, but none of them were correct. The task force also tried to compare tens of thousands of suspected photos with missing children one by one, but still couldn't find the children.
Under the traditional way of searching for people, many families of missing persons have paid a heavy price.
14 years ago, his son was taken away by the "Mei Yi" gang and then trafficked. Shen Junliang from Zhoukou, Henan has been searching for his son for 14 years. From the age of 28 to 42, he has hardly had a normal life. His wife has had mental problems due to the heavy blow and is currently receiving treatment. Since losing his child, he has never given up searching, even using the "dumbest" method to find his son.
[caption id="attachment_47223" align="aligncenter" width="540"] Shen Junliang posted missing person notices on the streets of Guangzhou, Dongguan, Zhuhai, and Shenzhen. Image source: Cover News[/caption]
After 15 years, Shen Junliang is worried that he won't be able to recognize his son, so he found a portrait expert to draw a sketch of his son. He hopes that with a portrait of his son as an adult, he won't miss his son if they cross paths.
[caption id="attachment_47224" align="aligncenter" width="540"] Shen Junliang's son, Shen Cong, in the portrait. Image source: Cover News[/caption]
However, after fifteen years, who knows what the child looks like now? Perhaps AI is breaking through this challenge.
According to industry insiders, AI search for missing persons mainly consists of three processes: pre-task, face feature extraction, and face comparison and retrieval. The photos of missing persons are used to extract facial features and enter them into the database system. At the same time, dynamic face monitoring is conducted at places such as train stations, airports, and shopping malls. If a passerby with a high similarity is found, the system will issue an alert, which usually requires manual verification.
As of August 30, 2019, on Baidu's AI search platform, users have initiated over 280,000 photo comparisons, and Baidu's AI technology has helped 7,619 missing persons reunite with their families.
Tencent's YouTu Lab, relying on years of accumulation in the field of computer vision and the accuracy of massive data retrieval, has exceeded 99.99% accuracy. So far, Tencent's YouTu has assisted police in anti-trafficking and missing persons cases in Fujian, Sichuan, and other provinces. Among them, the "Concerned About You" anti-missing platform of the Fujian Provincial Public Security Department has found over 1,091 missing persons; as of October 2018, the QQ Citywide Assistance program, which has integrated Tencent's YouTu "Sky Eye Search" function, has found over 600 people. Through years of accumulation in public welfare search, Tencent fully demonstrates the social value of technology and brings more possibilities to society through AI.
After Children's Day in June 2019, the Ministry of Public Security released the latest news. The "Reunion" system, developed by the Ministry of Public Security and Alibaba, has been online for 3 years. The system has received a total of 3,978 cases of missing children and has found 3,901 missing children, with a recovery rate of 98%. In addition to facial recognition, location-based and precise push methods have also been widely used. The information is pushed based on the missing location, and the targeted push range expands as the time of disappearance increases. Therefore, the "Reunion" system is also known as the nemesis of human traffickers.
With the support of technology, AI search for missing persons directly uses facial recognition, big data mining, and information push. The search is moving from the traditional 1.0 era to the 2.0 era. The new generation of search technology has made a qualitative leap compared to traditional search methods.
There are still several obstacles for AI search for missing persons to overcome
AI search for missing persons focuses on two key factors: the accuracy of facial recognition and the size of the facial database.
In the eyes of technical personnel, facial recognition requires a large amount of machine learning to make more accurate judgments, and the algorithm also needs a learning process, supplemented by a large amount of data, to be more accurate and effective. After all, the blurriness and lack of missing children's photo materials, as well as the significant changes in their facial features after years of disappearance, pose challenges to the algorithm.
Currently, first-tier and second-tier cities have a large number of surveillance devices, and the collection of relevant social aspects, such as ticket purchases, shopping mall information, internet usage, and hotel registration, is more easily monitored. However, human traffickers often operate in remote rural areas where surveillance is lacking, making it difficult to locate them.
Getting more people involved in AI search for missing persons and becoming the eyes of technology is one way to make up for this. So far, many people are still unaware of the new search methods.
Several industry insiders also call for the dissemination of relevant information to more people, especially rural friends, through precise flow push, so that everyone can join the camp of hunting down human traffickers and finding missing persons.
"Connecting existing data from all parties can improve the efficiency of finding people," an insider revealed, "the databases for finding missing persons, local surveillance videos, and social surveillance videos are scattered across different regions."
Tang Haipeng, a security expert from Tencent's Security Management Department, said, "Although this system is already mature, it will take a long time to promote and use it nationwide because it involves many coordination issues. We have coordinated several months for a case in one province."
Furthermore, another obstacle that hinders the development of AI search for missing persons is privacy and ethical issues.
The public is becoming increasingly sensitive to data privacy. On October 28th, Guo Bing, an associate professor at Zhejiang Sci-Tech University, sued a zoo in Hangzhou for changing the previous fingerprint entry authentication to facial recognition without user permission. In addition, the campus surveillance project and the campus monitoring incident caused by brain-computer interface headsets have also aroused public concern.
As a technology, facial recognition itself is not a problem, the key lies in how it is used. The same facial recognition technology, when used to establish a smart security system in kindergartens and primary schools to find lost children or elderly guardians, rarely causes controversy.
However, at the same time, the use of AI technology in educational settings requires further promotion of the rule of law, the establishment of relevant laws and regulations, and the assurance of legal compliance. This is the key to ensuring that technology is not abused, limiting the negative impact of technology, and truly realizing its role in promoting education.
Tang Haipeng previously revealed to the media, "There are significant ethical and moral challenges here, and we are under a lot of pressure... We can only be more cautious and try to avoid negative impacts as much as possible."
In a report titled "The Connotation and Compliance Challenges of Facial Recognition Information in Artificial Intelligence Series" by King & Wood Mallesons Research Institute, it is recommended that "enterprises should continue to comply with the principles of legality, legitimacy, and necessity when deploying and applying relevant technologies, conduct necessary personal information security impact assessments before deployment, and ensure that the processing of biometric information strictly complies with laws, regulations, and regulatory requirements."
Apple CEO Tim Cook said, "I'm not worried that machines will think like humans, I'm worried that humans will think like machines, without values, without sympathy, without reverence for the result."
Alphabet Q1 paid subscribers reached 350M, fueled by YouTube & Google One. YouTube ad revenue slightly missed at $9.88B amid Premium growth. Total Q1 revenue hit $109.9B, beating forecasts.
Anthropic releases Claude connectors for Adobe, Blender, and Ableton, enabling AI workflow automation and app control. It also sponsors the Blender Development Foundation.
Apple introduces an App Store model splitting 12-month plans into monthly installments. On iOS 26.4+, it lowers upfront costs without reducing total price, adding billing tracking.
Google launches Aadhaar digital IDs in Wallet for India, enabling secure QR/NFC verification. A phased partner rollout targets key sectors, while passport-based credentials expand globally.
YouTube tests "Ask YouTube," an AI tool blending text and video for complex queries. Initially for US Premium users, it supports interactive follow-ups, expanding Google's AI search.
Google grants US DoD classified AI access, joining OpenAI and xAI in defense deals amid Anthropic's dispute. Contractual safeguards against surveillance and lethal weapons remain unenforceable.
Mistral AI launches Workflows, a Temporal-powered orchestration platform for secure, Python-driven enterprise AI automation featuring human-in-the-loop approvals.