Future of digital marketing: AI changes rules of game

Professor of biochemistry and sci-fi author Isaac Asimov was one of the first to popularize artificial intelligence. Everybody knows his Three Laws of Robotics derived back in 1942. But the AI in the form described by Asimov does not exist yet. Today by AI we understand computer programs that can deal with certain tasks like humans or even better. Scientists and researchers divide them into weak and strong. The former can deal with narrow tasks while the latter are equal in intelligence to humans. Let’s see what type of AI is accessible now, what possibilities it gives to the digital market and where it is useless.

Expectation vs. reality

Shane Legg, a scientist studying machine learning at Google DeepMind, suggested that strong AI would emerge in 2020, followed by a decade of fast progress. Several years later, Mark Zuckerberg supported his prediction. In 2015, he announced that he would soon create an AI exceeding human capabilities. That was also promised to happen in 2020.

Despite the ambitious statements, however, we only have a rather weak AI at our disposal in 2023. And the list of tasks it can do for digital marketing and the society in general is still limited. But it does not mean that it is useless. AI takes up routine tasks and allows experts to focus on creating quality hypotheses and AI monitoring.

What can AI do right now?

Analyze big data or a company’s data, such as the market, rivals, consumer behavior and preferences, including using face and voice recognition (Cloud Vision, Cloud Natural Language, Cloud Video Intelligence).
Plan and launch marketing campaigns, automate and optimize them. Due to selecting the audience and keywords it allows for cutting advertising expenses.
Make translations and write code (Codex), create offers and texts (GPT), images (DALL-E, Midjourney) for marketing purposes, as well as select keywords for SEO promotion.
Communicate with potential clients and respond to users’ questions, select targeted content, etc.

However, as of now, none of these capabilities are implemented by themselves. To avoid mistakes, all AI-produced results are processed by humans.

What capacities digital marketing requires but does not have yet

To see what technological solutions are in demand with digital marketers, one should outline the main goal of marketing: to provide the clients with the best service possible. It means that the most desirable services are those that will help experts understand the psychology of the audience and predict what products the market needs, as well as when and how consumers would notice them.

Marketers currently receive audience information from AI, but the accuracy of this data is low. A weak AI can recognize a text, audio and video materials through which it analyses the demands of people, but cannot deal with their motives, impressions and emotions. AI also has difficulties with recognizing human speech. The message can be different depending on what words a person uses to form a sentence. The machine cannot sense the context and the tone.

Therefore, experts do need a strong AI – the one Legg and Zuckerberg spoke about – but it does not exist yet. The reason is simple: researchers still do not fully understand how a human brain works, so it is impossible to use its model to teach AI. Moreover, human brain interprets information gathered by the five senses: sight, sound, touch, taste, and smell. If the computer cannot have a human body, most of the data from the outside environment would be inaccessible to it.

What capabilities AI has but they are of no use

The development of artificial intelligence has led to creating products that allegedly serve people’s needs. For instance, ZERO10 created virtual fitting rooms AR Mirror for Tommy Hilfiger stores. Customers can have a try-on experience when shopping online, they only need to look at the screen with their photo-realistic avatars in a chosen outfit. However, many market giants such as Lamoda, ECCO and Wildberries failed to adopt the augmented reality technology. You can visit their websites and easily see that clothes are still showcased by models while shoppers try on items in traditional fitting rooms – and this is because we live in a real world and perceive it through all our senses. AI remains an amusing and expensive toy that is difficult to adapt to actual human needs.

This also applies to digital bloggers that brands use as virtual influencers, such as Shudu Gram, Lil Miquela, Imma, and others. Although having an impressive number of followers in social media, digital influencers are a short-lived trend: virtual bloggers are interesting only as a phenomenon, with audience aware that these fictional characters are created by a team of developers, artists and marketers. Such technologies can be entertaining for a while – yet, social media users keep returning to traditional blogs kept by real people.

Why not all Russian companies are keen to use AI services

In Russia, just like in any other country, major companies made use of AI first – such as banks, which have an immense amount of categorized data on users and transactions, as well as budget for testing and implementing AI technology. Successful solutions are normally pursued by large digital industry players. However, it is yet early to claim that AI has gained widespread use, with its slow facilitation and SMEs simply unaware of its capabilities.

Also, AI operates at the instigation of researchers and experts: they claim that an AI-based software program replacing one specialist requires two or three people to manage and monitor it. In 2023, the question as to where such competent workers can be found remains open, with methods of training new staff still to be developed and those available not commonly used due to distrust in AI in the society.

Other causes include:

the lack of technical infrastructure to support AI services;
small business owners rarely maintaining data in order, while AI requires ordered information to provide proper results;
people resisting change, which leads to any attempts to introduce AI in operations met with resistance from staff members and management team;
data privacy concerns surrounding AI and related questions regarding laws that do not regulate AI use in our country;
uncertain results: AI can be both helpful and useless, and it is totally unclear whether investments will pay off for SMEs, with such finances often unavailable.

Conclusions

The role of AI for society is immense – and it could become even more extensive in the long-term. Even weak AI offers plenty of opportunities for both creating products such as translation services, navigation, event planning and content recommendations, and promotion in the digital environment such as precise targeting and advertising campaigns, as well as personalized user communication tailored to their interests. But still, for now, the main job for AI in digital marketing involves solving routine and low-paid human tasks rather than becoming human. Although we do not know whether the latter is possible, a computer will certainly not equal human beings in the coming decades.