“AI ResurrectionShort VideoProduction, easy to fire the whole platform, public domain private domain double cash operation. This kind of video increases the powder fast, basically the small red book can increase more than 10,000 fans in a month, and the jittery voice can increase to 10,000 fans in a week if it is done well, and people who have this kind of demand are very willing to spend money to produce videos."
"Using AI to make a deceased loved one appear in front of you again, even if it's only for a short minute, can make up for the regrets that many people have in their hearts, so there should be quite a lot of people with this kind of demand as well, and there are very few people who know how to do this type of video right now, and there's relatively little competition."
"This type of video you can set your own price, anyway, my side to understand the peer pricing is basically 199 yuan a video, a video length of 3 minutes or less, you can make 1,000 yuan a day to pick up a 5 single."
…
Bao Xiaobai used AI to "resurrect" his daughter who died of illness, and then some netizens used AI technology to "resurrect" Coco Lee, Qiao Renliang and other deceased stars, and the recent fiery "AI resurrection" project has sparked a lot of controversy, but it has also made many people see the potential of this business. Although the project has sparked a lot of controversy, but also let many people see the development potential of this business.
From the use of a few dollars to tens of thousands of yuan order price to attract traffic, to a group of "resurrection teacher" through the course to share the industry bigger, and then apprentices in the mainstream short video platform to start the number to make money, "AI resurrection" of the commercialization of the road, has really gone on the right track of "as long as the payment can be customized"? "As long as the payment can be customized" on the right track?
How easy is it for AI to "resurrect" the dead?
How sophisticated is the "AI resurrection" technology now?
Only need to meet the following three requirements can be: First, to come up with a photo (including the front face) or a full body photo (including the front face), do not require high-definition, but you have to see the face clearly; Secondly, there is a 5 ~ 30 seconds long sound material, WeChat voice, audio recording can be used to reproduce the sound; Thirdly, you have to prepare a paragraph need to be through the generated loved ones to speak out of the words of the material.
How immature is the current "AI resurrection" technology?
Due to the amount of material and the cost of technology, there are currently two types of "AI resurrection" products that can achieve this effect.FirstKind is, the use of sound cloning, image processing and other technologies, directly let the characters in the photo open mouth "talk", that is, to generate a photo that can move; the second is, through the expression capture-driven AI face for private customization, with advanced AI technology to synthesize the character's real image and voice, you can carry out voice or video and other interactions. Unfortunately, these two processing methods can not get rid of the obvious "AI flavor".
Nevertheless.The "AI Resurrection" technology, which has both obvious advantages and disadvantages, has actually become a good business that many people have praised.
"Through advancedAI ToolsIn the past, the deceased loved ones could be 'resurrected' as if they had never left, with just a photo. They can speak and tell you of their thoughts and love in that familiar and intimate voice, as if turning back the clock and allowing love to travel through time and space. This is not only a remembrance of the past, but also a look into the future of technology."
"Using AI to make a deceased loved one appear again, even if it's only for a brief one-minute conversation, can make up for the regret inside."
"The purpose of doing the video is to make use of customers nostalgic feelings, so that friends who have lost their loved ones to see the dynamic effects again, to help them get back their memories. Dynamic old photos are not showy, helping people get back their emotions and memories is the most important thing."
AI Resurrection Tutorial on Idle Fish
As these generic advertising phrases suggest.The "emotional need" that such businesses market is the key to sustaining this business.
The relevant salesperson told Zinc Scale, "The people who have this kind of demand are usually people who have a very longing in their hearts, or their loved ones have just passed away and they have not seen the last of them, and there are also many people like Bao Xiaobai who sends the black hair to the black hair of the people with white hair, and their purchasing power is all very strong."
User comments under videos related to "AI Resurrection"
Therefore, even if many skeptics, these "loved ones" with the deceased's life image samples "cloned" out of the "AI resurrection" video of the actual performance effect is not how, but for the Its real audience, can already play a certain role in comfort, not to mention that many of the "AI resurrection" products on the market are not expensive, to convince them to spend money to experience it is not difficult.
Outrageous AI Resurrection Course:Was $980, now $6.88
During this time.The popularity of "AI resurrection" has led to an explosion in demand for such services, as well as the creation of assembly lines for "resurrecting" the dead.--From sending short videos to start the number, to the whole platform distribution to attract traffic, to open the window to sell goods, and finally through selling classes to radiate more people.
Previously, there has been media exposure of the "AI resurrection" course with the celebrity effect to advertise the inner workings of the course. According to the Rule of Law Daily report, an AI resurrection teacher apprenticeship program for different groups such as mothers, white-collar workers, students, etc., as long as you are interested in learning AI technology, entrepreneurship or side business, you can buy 980 yuan of apprenticeship courses. Its course includes an AI Resurrection Apprentice Manual, mastering basic technical principles and learning to use AI resurrection tools, and a complimentary face-swapping course worth 2,999 yuan.
It is understood that after taking the course, students can make talking videos of photos of the deceased and dub them with AI, and the program is priced at 198 yuan. The headquarters will be distributed every day, ranging from 5 to 50 number of orders, each completed an order can get 50 yuan share. Alternatively, those who want to learn to create more realistic videos of the deceased with a voice over that is very similar will need to pay $19,800. In addition, payment is direct, there is no contract, and the course is completed in 3 days.
Perhaps this kind of report is fruitful, zinc scale in Taobao, Xiaohongshu, idle fish and other platforms search found that the current "AI resurrection" part of the course commodities have been complained by users by the platform to be off the shelves processing, and stay to buy the option is not much, the price is basically in the few dollars to a few hundred dollars ranging.
Zinc scale respectively in Taobao and idle fish to buy a "AI resurrection" course, but in the receipt of the disk link to open, only to find that the two price tag is not the same commodity content is actually the same. The video tutorial in the project introduction, pre-preparation, hands-on video, the realization of the four aspects of the explanation, a full range of how to do a new white people should be able to earn this money.
Course content
This "original price of 980 yuan, the current price of only 6.88 yuan" course, basically can be regarded as a "foolproof" teaching template, as long as the students get the material can be based on the course content of the demonstration of the completion of the operation, but the focus of the generation of the video is still to be used! Three important AI tools: sound cloning platform Baidu Fei paddle, can generate dynamic video according to the photo Ali EMO, and AI intelligent production video tool!D-ID.
According to the course instructor, the actual cost of producing such videos is really only the cost of using theVideo Production ToolsThe money, such as D-ID currently supports a free trial of 14 days, want to use longer or experience more features, you need to give the corresponding package cost.
Different packages of D-ID correspond to different prices.
In other words.Even a person who is not the least bit tech savvy can catch up with this business opportunity if they know how to use the production tools associated with AI digital people.
SoraDriving the "seed-bursting" development of video tools
Since the tool is so important, one can't help but think of Sora, the Vincent video tool that triggered a "big earthquake" in the movie industry some time ago.
OpenAI explains on its website, "Sora is capable of generating complex scenes with multiple characters, specific types of motion, and accurate details of themes and backgrounds." Based on this period of testing, the imagination of generative AI has been fully opened when it comes to seeing what such large models are capable of.
In fact, as early as March and April last year, when the AI big model just started in the domestic "out of the circle" at the time, Baidu Wenxin Yiyin and other domestic big model one after another on the market, has been in the part of the AI digital entrepreneurs to help gradually verify its commercial value.
NJ, which had completed about 600 "AI Resurrection of Loved Ones" related demand orders in a 10-month periodsuperHeads, whose team leader, Zhang Zewei, said in a media interview thatsuperThe head team is based on Wenxin Yiyin such a large model to do voice, image and other training, "This is a non-standardized service, each customer can provide the deceased material, the demand for the effect is not the same, you need to go to do customization. Therefore, for large companies, they will pay more attention to standardized, generic technology and products, and will not invest in such a segmented track. And for some new entrants to the team, a lot of cognitive understanding of the industry and user needs is often not enough, and the technical reserves are not enough."
Throughout the domestic players, more than 15 companies have launched video generation tools, including both Baidu, Ali, Tencent, bytes and other 6 giants, but also includes Aishi technology, raw number of technology, Zhixiang future and other 9 start-ups. With the impact of Sora deep into all walks of life, the process of domestic iteration of related technology has also begun to accelerate.
On March 15, Tencent, Tsinghua University and the Hong Kong University of Science and Technology jointly launched the graphical video model "Follow-Your-Click". Users can click on the corresponding area and add prompt words to make specific objects move. The video shows that by selecting a bird, thunderbolt, person, rocket, etc., a short cue such as "turn your head", "laugh", "sad", "launch", etc. can be generated. A few seconds of video are generated based on short cues such as "turn head," "smile," "sad," "launch," and so on. In the video with the cue "sad", the character looks down and hides his face as he cries, and his movements are smooth.
On March 23rd, Aliyun Magic Hitch Community announced "ModelScope-Sora Open Source Program" to help explore and innovate Sora-like models in China with the power of open source. The program will provide a one-stop tool chain for the development of Sora-like models, including data processing tools, multimodal datasets, Sora-like basic models, and training and reasoning tools. At this stage, Magic Hitch has released the industryThe firstData-Juicer, an open-source multimodal data processing system, contains more than 100 highly efficient operators that can dramatically improve the efficiency and quality of video data processing.
"ModelScope-Sora Open Source Program."
In addition, there is news that byte jumping in the development of video generation model called Boximator, byte previously responded that this is a video generation field to control the movement of the object of the technical approach to the research project, with the leading foreign video generation model in the picture quality, fidelity and video length and so on there is a big gap.
While no further technical details are available at this time, it is certain that theAs the relevant technology iterates, AI industry practitioners will be ushered in to bring a number of considerable opportunitiesAfter all, the "AI resurrection" project may involve ethical and legal issues. After all, leaving aside the ethical and legal issues that may be involved in the "AI resurrection" project, the topic has continued to ferment and hot during this period of time, but also from the side of the generative AI technology to confirm the prospect of landing.