The development of AI (artificial intelligence) has obviously exceeded the imagination of ordinary people, and with it, AI data annotators have gradually become the largest group in the AI industry...
At the end of 2022, the large-scale language model ChatGPT released by artificial intelligence lab OpenAI attracted the global media, and in just 4 days, its number of users reached one million, and the number of registered users caused the server to be full for a while. Even Musk believes that "we are not far from an AI that is dangerously powerful."
In just 4 months, ChatGPT has gone through several iterations, and ChatGPT4 was released in April 2023, and at the same time, ChatGPT5 is also coming out.
The boom of artificial intelligence has made more and more people interested in it, and the "artificial" behind artificial intelligence has also surfaced, and the profession of AI data annotator has begun to be focused by public opinion.
AI data annotators, the "trainers" of artificial intelligence, they are the enablers of the AI industry and the "artificial" behind artificial intelligence. Behind the AI boom, there are billions of terabytes of data information passed through their annotation to AI, and finally let artificial intelligence achieve "intelligence".
In a high-tech development zone in Jinan, Shandong Province, team leader Dong Heming (pseudonym) sits in his office waiting for employees to send him the finished labeled data. In the office area separated from the office, dozens of data annotators are "box selection" of pictures that need to be marked, and the status of most front-line data annotators is no different from that of assembly line workers.
Annotating, framing, and tracing pictures is the daily work of AI data annotators, because the ontology of artificial intelligence is only a running program, and there is no human "eyes, ears, mouth, nose" to recognize the world, and all "intelligent" parts need AI data annotators to help this program achieve the purpose of recognition through the recognition and labeling of objects. And the more labels recognized, the more "intelligent" AI becomes.
When using a mobile phone to brush your face, completing payment, when you walk into a train station through face recognition, when you talk to your smart speaker and issue instructions to it, when you sit in a car and command the vehicle navigation system through voice, when you encounter danger due to deviating from the lane during driving, artificial intelligence intervenes in direction control. Behind these "technologies" that we are usually familiar with are the hard work of AI data annotators.
At Dong Heming's company, annotators have marked pictures with countless markers to create artificial intelligence "feed". Of course, today's Dong Heming does not have to do these front-line work himself, but every time he hears the sound of a mouse like a light raindrop in the office area, he still sighs very much about the time when he first entered the industry.
Dong Heming was originally an assembly line worker in an electronics factory in Kunshan, Jiangsu Province, after the end of 2019, the global epidemic caused the electronics factory to stop production, and in the following year, the factory hardly started, Dong Heming is also constantly trying new jobs, but failed again and again.
In 2021, Dong Heming learned about the work of AI data annotators from the Internet, and after applying and training, Dong Heming finally became an AI data annotator. Dong Heming's career as an AI data annotator began with the simplest and most mechanical entry. The work is boring, boring, simple, mechanical, but the salary is acceptable, 8 hours a day, can get a monthly salary of about 4,000, although the income is not much, but it is still precious to Dong Heming, who has not worked for more than a year.
Three months later, Dong Heming was entrusted with the work of data annotation, and the first job was to label the characters in the image, including age, gender, race, hair, expression, and submit the labeled data to AI for learning. At this time, Dong Heming's salary also rose to about 5,000, which is the norm for most AI data annotators.
Dong Heming often jokingly calls his work an "Internet migrant worker", the work is boring and repetitive, and ordinary people feel very new at first, but after three days, they can't stick to it. The reason is that although the entry threshold of this job is very low and the operation is simple, the requirements for accuracy are very high, and there are many positions to mark in a picture, and a slight deviation will make an error, resulting in inaccurate data packets, and even "rework" in serious cases.
In this way, after a group of colleagues could not survive to leave, Dong Heming still persevered. Because he is very clear that the work content of AI data annotators is actually to transmit the thoughts and cognition of "people" to AI, to put it bluntly, to use people as computers. Due to the tight time and heavy tasks sometimes, Dong Heming had to prepare Red Bull and American ginseng at his workstation, and the results of his hard work were good, and soon his monthly salary rose to more than 8,000.
Now, two years have passed, Dong Heming started from an ordinary trainee labeler, constantly honed his skills, understood the rules of the industry, and gradually reached the position of senior annotator, team leader and department head. The monthly salary has finally stabilized at 20,000.
At the same time, the group of data annotators is also expanding rapidly, and up to now, there are millions of data annotators across the country, mainly involving autonomous driving, face recognition, voice dialogue and other fields. The work of a large number of data annotators is also delegated to individuals in the form of crowdsourced tasks by large enterprises. But Dong Heming's company still has a large number of orders because it entered the industry earlier, and it is even busier.
Because after 2022, the domestic electric vehicle industry will make collective efforts, the verification of the L3 level of autonomous driving will continue, and the artificial intelligence developed by Internet giants will be online. The annotation data required for a large amount of artificial intelligence is simply astronomical, and it is estimated that in the five years from 2023 to 2028, at least 30 million data annotators will be needed in China. The gap is 10 times the number now.
Dong Heming believes that the data labeling industry will eventually make big waves, and they will be the last batch. Because they entered the industry early, they love this industry even more. Moreover, because the AI market has never stopped developing, AI is becoming the "entrance" for people to enter this society, which makes Dong Heming and other practitioners feel hopeful.
In the era of rapid development of science and technology, the application of various artificial intelligence has sprung up, and with it, the workplace is changeable. Dong Heming's success depends on continuous persistence and hard work, and when most people can't bear this boring and repetitive work, he completed his self-evolution and transformation. As for the new labelers, they have just stepped on the artificial intelligence train and still need to learn from their predecessors how to drive to a bright future.