On July 6th, visitors interacted with the Yumei humanoid intelligent service robot at Cloudminds booth. Xinhua News Agency (photo by Li Xinyi)
The opening song of the opening ceremony of this year's World Artificial Intelligence Conference is as cool as in previous years. Brilliant lighting, dancing 3D manipulator, vivid video pictures, combined with the performance of "Post-90s" erhu player and composer Zhu Yun, brought the audience an audio-visual feast that integrates various national musical instruments and rich national style.
It was not until the host "revealed" at the end of the song that the audience suddenly realized that the composition of the opening song, the performance of other national instruments, and even the production of videos and lights were all generated by artificial intelligence. Before you know it, generative artificial intelligence has entered people's daily lives.
Generative AI makes a difference
On July 7, the Yuan Radish AI chess robot played against the audience. Photo by Xin Mengchen, a reporter from Xinhua News Agency
Last year, generative artificial intelligence entered the public eye. Writing copywriting, writing essays, drawing illustrations, making videos... It seems that overnight, it has replaced the work of many human beings, making people happy and sad.
"Generation" means that artificial intelligence can not only do the previous discriminative work, but also form new productivity, in the long run, the scientific and technological revolution and industrial transformation triggered by "generation" are reshaping the world economy and affecting future life.
The variety of products "generated" by AI is accelerating its renovation. Just like the scene in the opening song, as one of the nine official "treasures of the town hall" of this year's World Artificial Intelligence Conference, Tencent Multimedia Lab's self-developed XMusic universal composition framework is based on generative artificial intelligence technology, supporting multi-modal content such as videos, pictures, text, tags, humming and other multi-modal content as input prompt words, which can generate high-quality music with controllable mood, genre, and rhythm.
This makes it much easier for people to create new music, whether it's a well-shot short video, a meaningful photo, or just a line of text that feels emotion.
Opposite the Tencent booth, the sound of shooting and shooting from the SenseShot smart basketball court in the SenseTime exhibition area attracted visitors to stop and participate from time to time. Relying on the five cameras set up on the court, the smart basketball court performs 3D modeling of the court through 2D key point extraction and 3D video motion capture and other technologies, and the system will automatically capture 29 key points of the face/human body and the basketball, and perform logical calculation and analysis on the players.
After the end of the exercise, the system can generate personal video highlights through the three-dimensional coordinates of people and balls in space, and analyze the action characteristics of people in different sports states, providing detailed data and statistics for players to help analyze and improve basketball technology.
"This artificial intelligence system can not only provide data statistics for amateur basketball enthusiasts to improve their technical level, but also provide technical statistics and tactical sharing for competitive sports competitions, helping professional basketball teams improve their technical and tactical level." Now we're reaching out to professional teams in the CBA and may provide them with solutions in the future. Wang Zhaohui, director of SenseTime's smart sports project, told reporters.
Big models empower thousands of industries
If Baidu Wenxin of last year's World Artificial Intelligence Conference was still unique as a large model, this year, more than 30 large models such as Ali Tongyi, Huawei Pangu, iFLYTEK Xinghuo, and SenseTime Ririxin all appeared, showing a full bloom of vitality.
"General artificial intelligence has become the hottest topic in our human society in this period." Ken Hu, Huawei's rotating chairman, said at the opening ceremony, "While Huawei is deeply cultivating computing power, it must also do a good job, that is, let artificial intelligence truly enter thousands of industries and serve scientific research." On the one hand, we must continue to improve the ability of general large models, and on the other hand, we must build industry models on this basis. ”
As the winner of the SAIL Award at this year's World Artificial Intelligence Conference, XtalPi's intelligent automated drug discovery platform uses AI and robotics to achieve a high degree of automation and intelligence in the innovative drug discovery process, including the application of professional models.
They took the lead in applying GPT-like technology to the research and development of macromolecule drugs, and independently developed the next-generation macromolecule drug-generating AI model ProteinGPT, which has good applications in scenarios such as antibody discovery, antibody modification, antigen design, and protein de novo design. As the world's first large-molecule drug-generating AI model that is implemented in all scenarios in a real drug development pipeline, ProteinGPT is also likened to a biological version of generative artificial intelligence in the industry.
Generative AI is a double-edged sword
On July 7, visitors visited the AIGC Art Exhibition with the theme of "Symphony". Photo by Xin Mengchen, a reporter from Xinhua News Agency
"Generative artificial intelligence is a double-edged sword, which not only benefits mankind with rapid development, but also raises new questions for mankind in terms of ethics, social governance, laws and regulations." According to Cui Yadong, president of the Shanghai Law Society, generative artificial intelligence currently has a series of problems to be solved in data security, intellectual property rights, application specifications and related laws.
"From a legal point of view, AI may generate false information, which can lead to infringement issues, which requires us to make the necessary legal response." Wang Liming, a professor at Chinese Minmin University, said that we should pay attention to regulation and also encourage innovation. "How to leave the necessary space for innovation is a value orientation that we must consider when determining infringement."
For the future of deep artificial intelligence, Tesla CEO Elon Musk is also a little worried. "Such superintelligence has great abilities and is more powerful than humans, which also represents a risk and a hidden worry." Musk said at the opening ceremony of the conference, "It may have a very positive future, but there is also a probability of causing some negative events, and we must do our best to ensure that these negative futures do not happen, and a positive future will happen." ”