laitimes

AIGC rolls up to the gaming industry, whose jobs are being replaced?

Interface News Reporter | Peng Xin Xiao Fang

Interface News Editor | Wen Shuqi

Guo Xin, who is in charge of the company's art asset management, is arguing with outsourcing companies significantly less often because he has begun to use AI drawing.

Previously, when using the outsourcing team, he often asked the outsourcing to return to the picture again because the details of the original painting were wrong and the theme of the game was not very suitable. "Maybe change it three or four times, and the other party will not be happy."

Now, with the popularity of various AI drawing software, Guo Xin has begun to try to use AI tools to generate pictures, eliminating the cost of communicating with art outsourcing companies.

The industry is already showing signs of change. According to Zhang Gang, an original artist of an art outsourcing company, a large number of original artists have been laid off due to insufficient workload due to the reduction in demand from Party A and the fact that his company has also begun to use AI to complete the creation of some original painting details.

Although AI painting is not perfect, Guo Xin, as party A, also has his own considerations for using it to replace outsourcing: art asset outsourcing is calculated per person-day, the price starts at least 1,000 yuan per person-day, and repeated modifications mean that the cost of an original painting is higher.

At present, the replacement of the original artist by AI painting is still an example, and it is only regarded by many people as a hot topic after tea. But what most people really care about is whether AI will actually replace human jobs in large numbers in a few years' time.

The shock starts with standardization efforts

It's no surprise that game outsourcing jobs are among the first to be impacted by AI.

Outsourcing is a well-established practice in the industrialization of games. By outsourcing the development of art, music, translation, etc., to reduce development difficulty and save costs, game companies form a stable symbiotic relationship between the two parties by outsourcing stable labor and standardized services.

Painting and translation are the positions most affected by AI in this wave of AIGC technology change.

In fact, translation was hit by AI earlier than painting, and many practitioners have "laid flat". Gao Yang, a veteran translator who has been engaged in game localization translation for many years, has realized that there is vicious competition in translation outsourcing companies. "Some companies in the industry maliciously bid in order to grab orders, and some people even proposed 'power generation for love' for free translation, and outsourcing companies took projects without making money, and the result was to use machine translation to fool Party A."

Under the vicious competition, the most affected are small translation studios, which mainly undertake game text translation that does not require so high accuracy and does not have additional liability risks.

All the shocks eventually fall on the individuals doing the work. Gao Yang lamented that he chose the translation profession in his early years to work freely and flexibly, and to avoid the age ceiling, but those who entered the industry as translators again in 2020 are not so lucky.

And the sense of crisis among art practitioners has just begun.

In the past three years, the game industry has invested heavily in improving the picture quality of its products, and has hired artists with high salaries. Especially in 2021, the salary increase of 1-2 times of art personnel is more common, and some people have increased by more than 3 times, and some practitioners describe that "it is outrageous to open a salary in a large company as if it does not cost money".

But now, some artists have suffered salary cuts, and some can't find jobs at all. For example, domestic mobile games that tend to develop card unboxing krypton gold require a large number of 2D artists to produce art content, but after the rapid rise of AIGC, it first solved the scale, standardization and industrialization of 2D picture production, improved production efficiency, and also impacted the work of artists.

Liu Yu, a game art practitioner, told Interface News that after using AI-assisted design, the refinement of the character design draft only needs to be manually produced by hand, and the design of pattern details, the refinement of the picture and other repetitive labor can be completed by AI. Finally, if you repair it manually, the degree of completion is considerable.

Liu Yu's company has introduced AI art, which can quickly produce dozens of pictures at a time, increasing production efficiency by more than 50%.

In 2D art, the advantage of the original artist lies in creativity, and he can participate in the initial development of the IP image, which is considered by the industry to be temporarily not affected by this round of AI, but practitioners still have a sense of crisis.

Liang Qiwei, founder of the game company Beijing Lingyoufang, said on his social media that AI is destroying humanity's last line of defense little by little. He sent out several AI drawings on social networking sites showing that by putting forward requirements for ChatGPT, and then referencing the concrete setting description of ChatGPT's output to Midjourney, the concept map of his final output was relatively close to his idea prototype.

Most practitioners are worried that the rise of tools such as ChatGPT will eat up a large part of the standardization work such as translation and painting that was solved by humans in the past, leaving only a small number of problems that cannot be solved by AI to be handed over to humans. Although both translation and painting have some complex links that cannot be replaced by AI, it is a fact that the industry has been hit by the market, resulting in a reduction in manpower demand.

"No one has done anything wrong, the main thing is that times have changed." Gao Yang said to Interface News.

AI changes game making

AI brings a sense of crisis to game practitioners because the iteration of technology is subverting game practitioners' expectations of AI innovation capabilities.

In October last year, Liang Qiwei bluntly said on social media that in the future, "drawing hands" means that only need to draw AI pictures, implying that AI drawing tools still have flaws in human hand drawing. Six months later, he changed his tune because he found that the latest Midjourney drawing model had a hand accuracy rate of nearly 50 percent.

The game industry has a long chain and a high degree of industrialization, covering many types of positions such as art, copywriting, translation, programming, marketing and promotion, and different positions can be subdivided many times on this basis to form a complete chain. The commercial fields involved in AIGC include media entertainment, copywriting, data analysis, programming, etc.

After seeing the latest iteration of ChatGPT's model GPT-4, the founder of a game company judged Interface News that the future information retrieval, collection, and induction related positions were unnecessary. What's more, when gameplay and AI are combined, the cycle for creating blockbuster games will be shortened, "a change that could happen in about two years." ”

The company considered importing AIGC text for a mobile game project under development last year, but the attempt did not go well. In terms of corpus, cost, and text quality, AIGC text is more difficult to apply than expected, but its founder pointed out that perhaps AIGC will be more helpful to other games with low text quality requirements, after all, it can produce a very large amount of text content, which may cause qualitative changes.

Other game companies are starting to use AIGC to generate copy on creatives. Wang Wei, general manager of mobile attribution and marketing analytics platform Appsflyer China, said that some practitioners told ChatGPT what kind of game it is, and asked it to generate copywriting that is in line with both game tonality and Christmas tone, and ChatGPT's ten copy results are quite good, reaching the medium level of the copywriting team.

Many game companies can use AI to generate marketing materials to save costs. These AI batch drawings do not need to be creative, and are only used for the "buying volume" of game marketing. This has a great appeal for game companies: games are launched in the face of fierce competition for buying volume, and game manufacturers have an incentive to control costs.

The potential of AI also lies in helping game developers optimize their workflows. Alexis Rolland, director of La Forge, an internal research center of French game company Ubisoft, told Interface News that Ubisoft has been actively researching solutions based on AI and machine learning, and many of the research and development has been put into production pipelines. For example, Far Cry 6 uses an AI tool called Choreograph to use the technology of matching movements to generate more natural animations, and Rainbow Six: Siege also plans to provide AI opponents to improve the player experience.

He explained that AI can automate tasks that are simply repetitive at work, allowing developers to explore and experiment more, create new content faster, and have full control over the entire creative process.

"We did an internal analysis to evaluate the use of AIGC in the gaming industry." The relevant person in charge of the game company told Interface News that he concluded that whether it is used to inspire ideas, find gameplay ideas, or generate world view settings in the design stage, including code optimization assistance in the development stage, and promotion materials in the operation stage, AIGC is cheap and sufficient, brain holes are quickly visualized, and "basic exercises" are proficient, and have good practicality.

More directly, AIGC can play a role in all stages of game production and distribution, from the creation of an original art to the entire production process of the game - AIGC subverts the entire game production and distribution industry chain, and ultimately affects the jobs in every link of the industry chain.

Uncertain future

With AIGC's strides, more and more game industry practitioners are facing confusion: how to survive in the future?

In fact, this is not a problem faced by the gaming industry alone - in the context of the GPT boom, the whole world seems to be swallowed by the volcanic ash of AI.

The White House's "Artificial Intelligence, Automation and the Economy" report said that automation will cause millions of workers to lose their jobs, widening the economic divide between social classes in the United States. In other words, in the near future, a large number of human jobs will be automated, and a large number of low-skilled workers will face unemployment. For example, the once common manual customer service is being largely replaced by AI.

In the AI era, game outsourcing in the semi-marginal area is not enough to climb to the top of the food chain to directly obtain huge profits, and may not be able to obtain new positions in the drastic changes in the system.

What will become of human life when more people take their jobs from AI?

Edward Castronova, a professor at Indiana University in the United States, boldly predicted in a white paper published in 2016 that within 20 years, being hired to play games will become an important source of income for low-skilled workers who have lost their jobs in the wave of automation. He argues that if game companies want to retain rich players in the future, they will have to pay for a large group of poor players to accompany them, which he calls the "Play-for-hire" model.

But some practitioners remain cautiously optimistic. Wang Wei said that based on AI, the current market has produced new occupations such as artificial intelligence trainers. AI can't automatically reach a very high level, it requires someone to ask the right question and give the right feedback. At this stage it replaced some jobs and produced some new ones.

In the recruitment market, the application of AI technology is driving the creation of new jobs in the industry. From the perspective of job segmentation, positions such as AI original painters and AI art designers have emerged to cater to industry trends. The requirements of such positions usually include understanding AI art-related tools and platforms, refining key words, generating benchmark original paintings, and refining and adjusting AI images.

Still others believe that AI cannot replace artists. Artistic creation with profound significance and great influence is inseparable from the artistic vision and perspective of human artists, and AI is only a tool in the artist's toolbox, and its existence is to meet the needs of artists.

While AI is giving more and more people a sense of crisis, the reality isn't as bad as Edward Castronova predicted. Some industries are still looking for dividends under the onslaught of AI technology, such as the game industry, which was the first to be hit.

The hot ChatGPT technology opens up design ideas for game manufacturers. NetEase announced the implementation of ChatGPT in the mobile game "Against the Water and Cold", allowing NPCs (non-player characters) to freely generate dialogues with players, and independently give logical behavioral feedback based on the content of the dialogue, that is, open dialogue, to enhance the sense of substitution.

CMGE has been connected to Baidu's "Wenxin Yiyan", testing the intelligent interactive function of NPCs in the game for their own "Immortal Sword World", ideally, NPCs can obtain independent character story backgrounds and personal personalities, and freely talk with players, NPCs can respond according to the set personality, which in turn affects the subsequent plot development of the game.

At present, most of the people who come into contact with AIGC technology are still in the stage of playing with new technology experience, and they are trying more possibilities brought by technological progress.

"Of all the internal groups in the company at the moment, the art exchange group is the most lively every day." The person in charge of the aforementioned palm fun said that at present, art colleagues like to share AI "creation" works in the group, show creativity, and discuss application in games, and everyone is very excited about the new toy.

Instead of worrying about a future that no one can accurately predict, enjoy the new experiences that technology brings.

(At the request of the interviewee, Guo Xin and Liu Yu are pseudonyms)

Read on