How AI will influence the development of creative professions

With the development of generative neural networks, workers in creative professions became concerned about their future fate. These include announcers and voice actors, artists and graphic designers, and employees of large IT companies. Kommersant looked into how neural networks can harm or, conversely, help in these professions, whether an economic effect from their use is possible, and whether AI can replace a living person.

Deputy Prime Minister Dmitry Chernyshenko said at the Russia exhibition and forum in March that a third of Russian companies are already using artificial intelligence. According to him, a serious bet is being made on AI as the main driver of technological independence in the next few years. But still, the neural network will not be able to replace the majority of jobs right now in cost-effective ways, says the report. research Massachusetts Institute of Technology (MIT).

Still, voice-over artists, for example, have reported cases of their voices being stolen and fear a decline in income as technology spreads. At the beginning of 2023, the Union of Russian Speakers demanded the establishment of special regulation for the synthesis of the human voice using artificial intelligence technologies.

The most obvious risk is that live actors will be left without work, says voice actor and localization director Ivan Zharkov. In his opinion, a layer of types of voice acting will be formed that will be much cheaper to do using AI. But neural networks will definitely not replace a living actor, he believes. “A neural network can do high-quality work – pronounce all the letters, etc., but now it’s just funny to listen to: incorrect stress, illogical sentence construction – computer synthesis, dry, soulless,” notes Mr. Zharkov.

Questions also arise with painting. For the last two or three years, artists and graphic designers have been suing the creators of neural networks for paintings drawn by AI in their style. Thus, in January of this year, the British The Guardian wrote that more than 16 thousand artists filed a lawsuit against a number of companies that create images based on the text they suggested, including Midjourney, Stability AI, Runway AI and DeviantArt. The neural networks generated pictures very similar to the original works of the artists.

The claims of the artists whose works were used as examples for training the neural network are justified and legitimate, says art director of the FFStudio design studio Fedor Filimonov. In his opinion, replacing artistic professions with artificial intelligence is a bad idea: neural networks operate on what has already been invented by humanity, and new thought is always required in this area. “The synthetics are visible. Some simple tasks, such as drawing something simple, can most likely be handled by the AI. But only an artist can draw a work with meaning, spirit and individuality,” says Fyodor Filimonov. However, he does not rule out that there are risks for those who draw worse than AI – they will not be able to compete with it and will lose their jobs.

Artists and graphic designers are skeptical about the results that a neural network produces.

And if they use it, then only for working with text. But as for graphic work itself, it is easier to give this task to an employee. “For a very long time you set the task of what should be drawn, and then you spend a long time correcting it. It’s easier for an artist to say: “Draw.” And he will definitely draw what is needed,” notes Fyodor Filimonov.

New technologies also influence the IT sector itself. Against the backdrop of the active development of generative neural networks, large foreign IT companies even began to cut staff due to the fact that they were successfully coping with the tasks of programmers. However, MWS Backend developer Vladislav Ramus most likely does not see the risks of using AI in the work of IT specialists. Neural networks, in his opinion, help to minimize routine, monotonous, template work: “There is an opportunity to spend less resources on tasks that are not about creativity, but about accuracy in execution. In addition, thanks to neural networks it is easier to obtain the necessary information and understand something new. This is a very powerful tool if you figure out how to use it,” the expert notes. The most popular tools are ChatGPT and Yandex GPT chatbots. Some also use the AI ​​assistant Copilot, which allows you to write, review and refine code.

Is AI as scary as it is made out to be?

In the near future, neural networks are unlikely to significantly change the labor market; the skill of using chatbots will become a competitive advantage, many experts believe. Thus, the chief architect of artificial intelligence systems of the research center in the field of AI in the field of “Transport and Logistics” NRNU MEPhI Roman Dushkin reminds that, first of all, AI is a tool created by people to increase productivity and efficiency in the field of mental work. “AI systems will not replace people, they will become instrumental assistants for every person. They will only replace those who are not ready to develop new skills,” he emphasizes.

According to Ivan Zharkov, the very presence of a neural network is not as scary as its unfair use. He told Kommersant that one company is using his voice, given to the character, for commercial purposes without consent. She offered Ivan cooperation for royalties. There followed a refusal and a demand to remove the voice from the neural network. “And then they changed their shoes and said: “This is not your voice, this is a parody.” Of course, they couldn’t provide me with a parodist,” said Mr. Zharkov.

Mikhail Khokholkov, head of the Media Law practice at the Intellect law firm, notes that such unscrupulous users, if a lawsuit is filed against them and they cannot provide a parodist, will lose the court – it will be recognized that personal non-property rights of a person have been violated (illegal use of voice) and direct damage to business reputation has been caused.

How to protect yourself today? In the case of personal rights, using the Personal Data Law.

“Any biometrics is allowed only on the basis of the person’s written consent – an agreement is concluded, which will indicate the purposes of processing and their scope,” notes Mr. Khokholkov. Voice assistants, for example, work on this principle. In the case of copyright, according to Mikhail Khokholkov, the most correct thing now is to create user agreements for services. For example, the generative model of Sberbank Kandinsky indicates that the image was produced by it.

Question of authorship

There is currently a debate among lawyers: who owns the result created using AI—the author of the software that made it, or the person who wrote the request? Mr. Khokholkov notes that in the current version of the Civil Code of the Russian Federation, the author can only be the individual whose creative work created it. And in such a paradigm, according to the expert, the issue of authorship, until appropriate changes are made to the legislation, can only be regulated by agreement.

According to Roman Dushkin, every person can use AI for good today. For example, create your own personalized systems. Mr. Dushkin implemented such a system in his company – the chatbot “Aurelia” (she chose the name for herself). According to the expert, it allows each employee to solve routine tasks. “Employees can go to the chatbot in Telegram and make a request. For example, “compile a semantic network of concepts for such and such a subject area.” The system will compile and define each concept with references to regulatory documents of the Russian Federation,” he says.

Also, a chatbot created on the basis of Aurelia has been introduced into the educational process of NRNU MEPhI and helps students study in Roman Dushkin’s course on cellular automata. “The chatbot has a complete knowledge base on this topic, my lectures, methodological recommendations for conducting laboratory work, organizational and technical issues related to the course, links to additional videos and literature,” says the expert. And what’s important is that a student can write her a request for a solution to some problem, and the large language model does not invent it, but will contact another system that will solve it clearly according to the algorithm.

But still, experts note that right now it’s hardly worth expecting an economic effect from the use of AI.

Vladislav Ramus says that neural networks have not yet become a standard and many specialists do not use them. In the future, when this technology has wider application, the development and launch of IT products will accelerate, he notes.

According to Ivan Zharkov, even in the future, without harming the profession, AI can be used only in one case: when a sound engineer needs an urgent edit, and the actor cannot come to the studio, the neural network will fake the voice of the right person, but again only with his consent.

Fedor Filimonov believes that, most likely, over time, people will be able to distinguish works of art from works created by AI, and in general there may be a reorientation towards things made by people. “A person’s eyes will already be trained: “Oh, is this artificial intelligence? You can skip it.” The feeling that something will be valued as a hand-made thing, something made by hand, is not the same as something made in a factory,” he notes.

Pelageya Owl