Latest Post

IDC report: The domestic game cloud market is stable and improving, and Tencent Cloud once again leads the way in usage scale PayerMax attended Seamless Middle East: Matrix services help overseas companies achieve safe and rapid business development

[GameLookspecialarticleunauthorizedreproductionisnotallowed!】

GameLook report / On July 6, the 2023 World Artificial Intelligence Conference (WAIC), co-sponsored by the National Development and Reform Commission, the Ministry of Industry and Information Technology and other government agencies, associations and the Shanghai Municipal People’s Government, officially opened.

Since its establishment in 2028, WAIC has attracted a large number of domestic and foreign politicians, experts, scholars, entrepreneurs and other people from all walks of life to participate, and has now become one of the most important industry events in the global AI field.

In the past year, AI has demonstrated impressive strength in various industries including games, finance, and content creation, and the overall AI capabilities are still improving, which also makes this year’s WAIC 2023 more worthy of attention .

According to the information on the official website of WAIC, this year’s conference attracted more than 1,300 guests, more than 400 exhibitors, more than 50,000 square meters of offline exhibition area, and hundreds of forum activities will be held.

When discussing AI development, games were rarely mentioned in the main conference

In addition to the opening ceremony of WAIC yesterday, the plenary session of the WAIC Industry Development Forum in the afternoon was the most important guest lineup, including AMD, Stability AI, Midjourney, and executives from companies and organizations such as Qualcomm, Tencent, JD.com, Baidu, and CITIC Institute. Gathered with experts and scholars, with “AI breakthrough”, “industrial butterfly change” and “future kinetic energy” as the three main lines, discussed a series of issues at the forefront of the current AI field.

However, GameLook is a little bit unexpected. As the most active industry in exploring AI in the past year, the presence of games in the plenary meeting is extremely thin. Only two companies, Tencent and Qualcomm, mentioned in their speeches that ” game”.

Amazon Cloud Technology actually mentioned it, but combined with the context of Shirley Marcus, director of Amazon Cloud Technology’s Generative AI Product Research Institute, the “game” is more inclined to traditional sports events than video games.

In fact, when JD.com and Bank of Communications talked about AI models and application scenarios, they did not involve games, but focused on logistics, finance and other fields. GameLook is not incomprehensible, after all, the technology industry has specialization.

But some other participating companies, such as Stability AI, Midjourney, etc., are somewhat strange not to mention a word, considering that their products are widely used in games and other related fields.

Insert a cold knowledge, according to MidjourneyMidjourney founder David Holtz cast spells, the name “Midjourney” actually comes from the concept of “middle way” in Chinese Taoism.

Combined with the field of AI painting that Midjourney is in, ChatGPT believes that the name can be understood as “In the process of creation, humans can use AI technology to innovate, but at the same time retain the uniqueness and individuality of art, and achieve a balance between art and technology. “

Game design generated by Midjourney

Back to the topic, we must know that the current game industry in China, whether it is regulatory or public opinion risks, has clearly shown signs of improvement, and avoiding risks is obviously not the reason.

Although it is not difficult to see some trends in the current AI field from the guest lineup of the plenary session of the WAIC Industry Development Forum. For example, the influence and strength of vertical models are constantly being demonstrated, and AI is no longer just an era of large models.

Large models refer to ANN models with a very large number of parameters, while vertical models refer to ANN models that target a specific domain or industry, but utilize domain-specific corpus for pre-training or fine-tuning.

The best example is that Stability AI and Midjourney can be invited to participate with AMD.

Considering that games will be at the forefront of technology and culture in the future, they must also be the key scenarios for AI application and exploration. At present, the large models that are trying to tackle AGI are still timidly drawing a line from the game industry, which seems somewhat inappropriate. wise.

Generative AI becomes an imagination engine, openness and closure still need to be weighed

Although there were very few games at the conference, leaders from different industries expressed their views and predictions on artificial intelligence. Their speeches are connected, and their impact on the development trend and potential of artificial intelligence is still worthy of attention.

In addition to JD.com, Baidu and other participating companies centering on their own big models such as Wenxin Yiyan, explaining their potential assistance to the company’s various businesses, AMD, Stability AI, etc., predicted and judged the future development of AI from a more grand perspective. possibility.

Image source: AMD official website

Among them, Su Zifeng, as chairman and CEO of AMD, expressed concern about the rapid adoption of technology. She pointed out that the rapid development of technology makes artificial intelligence play an important role in scientific research in various fields. Especially in the field of health care, the widespread application of artificial intelligence can help doctors make better diagnoses and accelerate disease prevention research. Su Zifeng believes that the application of artificial intelligence will achieve a significant acceleration in the field of healthcare.

Emad Mostak, the founder and CEO of Stability AI, focused his speech on the importance of open source. Openness, he says, is key to advancing human potential, especially in the age of big models. He believes that black box models will not be enough to turn data into intelligence and use it in parallel, people need to be able to see how models are generated and decisions are constructed. Emad Mostak believes that the joint cooperation of open and closed solutions will help create intelligent systems that can absorb information and realize human potential.

As for Jim Casey, chief business officer of Qualcomm Technologies, it is mainly around the combination of 5G and artificial intelligence. He pointed out that 5G is the key infrastructure for realizing a smarter world, promoting the application of artificial intelligence in fields other than smartphones, and promoting the large-scale development of intelligence. Jim Casey emphasized the application potential of artificial intelligence in smartphones, industry, automobiles and other fields, as well as its positive impact on underdeveloped areas, medical care, education and other fields.

Midjourney founder David Holtz came up with the concept of an engine of imagination. He believes that artificial intelligence can become an engine of imagination. Through the application of generative AI, people can imagine all possibilities before creating, thereby enhancing the original power of creativity. He emphasized the significance of generative AI for artists, designers and media producers, providing infinite possibilities for creative space.

By synthesizing these speeches, we can see the wide application and great potential of artificial intelligence in different fields. Artificial intelligence can not only accelerate scientific research and medical innovation, but also improve user experience and quality of life. However, the development of artificial intelligence also faces some challenges and considerations.

One such challenge is the issue of data privacy and security. Su Zifeng pointed out that the privacy protection of personal data is an important issue, especially for entrepreneurs and individuals, it is very important to have a private space to view personal data. While promoting the application of technology, we need to consider matters related to intellectual property rights and confidentiality to ensure the legality and security of data.

Another challenge is the balance between open and closed AI. Emad Mostak emphasized the importance of openness, arguing that open source models can advance human potential. However, at the same time, we also need to consider how to protect intellectual property and business secrets. In future developments, open and closed solutions will need to work together to create stronger, more secure intelligent systems.

In addition, the development of artificial intelligence requires interdisciplinary cooperation. Su Zifeng pointed out that the next-generation chip design needs to cross the boundaries of hardware, software and algorithms, and needs to truly understand the usage, deployment and application scenarios of the entire system. This interdisciplinary collaboration will help to better design products and provide users with a better experience.

To sum up, from medical care, scientific research to smartphones, industry, games and other fields, while the application of artificial intelligence is broadly expanding related industries, it also raises many challenges for us, such as data privacy, the balance between openness and closure, and interdisciplinary education. aspects of the challenge. Only by addressing these challenges will we be better able to explore the future of artificial intelligence.

Games may be the first biggest scene of AGI, AI companies need to pay attention

Although looking at the entire WAIC 2023, the presence of the game industry in WAIC is still very strong, including Tencent, Netease, Mihayou and other domestic leading game companies have participated in this conference, and Tencent will also host a famous event on July 8. It is called “Game AI Application and Game Technology Forum”, and the organizer of WAIC’s Game AI Metaverse sub-forum last year was MiHoYo.

Although they did not appear in the main forum, current leaders in the AI ​​field such as Nvidia, OpenAI, and Microsoft and the popular fried chicken revelations are also at the WAIC conference. For example, Microsoft also has a separate sub-forum held at the same time this year, with more than 30 large models at home and abroad. Also gathered at WAIC 2023.

Image source: WAIC official website

But as a game media, GameLook still wants to emphasize the importance of games in the process of AI development to AGI (General Artificial Intelligence).

The so-called AGI actually refers to the concept of a machine that can integrate countless reasoning abilities close to human level. AGI is different from artificial intelligence in the narrow sense, which is like the technical level of some applications in the field of artificial intelligence at present, and can complete specific tasks, such as image recognition, natural language processing or chess playing.

AGI, on the other hand, aims to be able to learn from knowledge in any domain and apply it to any problem, just like humans.

As for why the game GameLook firmly believes that games will play a key role in the emergence of AGI.

First, a game is a simulated environment that can provide AI systems with rich and varied data to learn from. Games can also provide a variety of complex challenges to better train the comprehensive capabilities of AI large-scale model reasoning, planning, creativity, and adaptation.

Second, games are interesting and attractive to users and the market, which can motivate users to actively explore and use AI tools, provide enterprises with a steady stream of samples, data and scenarios, and strengthen related AI skills.

As a concrete expression of human culture, games can also help AI systems better understand and interact with humans, and become a platform for cooperation and competition between humans and artificial intelligence systems, which can promote mutual learning and cooperation between humans and AI. cooperate.

Games can also be an important source of inspiration and creativity for AI systems, as they can expose AI systems to new concepts, stories, characters, aesthetics, etc.

Finally, games can also be a tool for testing and evaluating the safety and ethics of AI systems, avoiding unnecessary harm or suffering to real people.

Of course, the more specific role of games in the development of AGI remains to be seen, and even speaking realistically, the author cannot guarantee that games will be a sufficient or necessary condition for the emergence of AGI.

However, at present, commercial products including the mobile game “Naishuihan” have boldly embraced AI tools, while many traditional companies are still worried about privacy, data, accuracy and other issues, and dare not use AI, including many giants in the industry.

Although games are one of many tools, and may not even be the only or the best way to realize AGI in the future, it can help us create AGI that benefits all mankind. In related fields, at least what society can give is to Games with more emphasis and attention.

If reproduced, please indicate the source: http://www.gamelook.com.cn/2023/07/522788

Leave a Reply

Your email address will not be published. Required fields are marked *