Selected analytical information

Have you ever wondered how AI will change the course of history?


Generative AI creates value now. The speed of scaling by properly mobilizing the organization and focusing on value will be far more important than the speed of piloting.

AI has infiltrated our lives gradually, through everything from the technology that powers our smartphones to the autonomous driving features on cars to the tools retailers use to surprise and delight consumers. As a result, its progress has been almost invisible. 

Generative AI systems fall under the broad category of machine learning, and here’s how one such system, ChatGPT, describes what it can do:

Ready to take your creativity to the next level? Look no further than generative AI! This nifty form of machine learning allows computers to generate all sorts of new and exciting material, from music and art to entire virtual worlds. And it’s not just for fun-but generative AI has many practical applications, such as creating new products and optimizing business processes. So why wait? Unleash the power of generative AI and see what amazing creations you can come up with!

Did anything in this paragraph strike you? Probably not. The grammar is perfect, the tone works, and the narrative flows.

That’s why ChatGPTGPT means a generative pre-trained transformer – which now gets so much attention. It is a free chatbot that can give an answer to almost any question asked of it. Developed by OpenAI and released for testing by the general public in November 2022, it is already considered the best AI chatbot. And it’s popular, too: more than a million people have signed up to use it in just five days. Starry-eyed fans have posted examples of chatbot computer code creation, college-level essays, poems and even halfway decent jokes. Others, among a wide range of people who make their living creating content, from advertising copywriters to full-time professors, are quaking in their boots.

                                          Exhibit 1

Users don’t need a degree in machine learning to interact with or derive value from AI; almost anyone who can ask questions can use it. And, as with other breakthrough technologies such as the personal computer or iPhone, a single generative AI platform can create multiple applications for audiences of any age or education level and anywhere with Internet access.


How generative AI differs from other types of AI

As the name implies, the main way in which generative AI differs from previous forms of AI or analytics is that it can effectively generate new content, often in “unstructured” forms ( such as written text or images) that are not naturally represented in tables with rows and columns (see panel below“ Dictionary ” for a list of terms related to generative AI).


As the name implies, the main way in which generative AI

differs from previous forms of AI or analytics in that it can effectively generate new content, often in “unstructured” forms ( e.g., written text or images) that are not naturally represented in tables with rows and columns (see the “Glossary” sidebar for a list of terms related to generative AI).

The base model that allows generative AI to work is called the base model. Transformers are key components of base models – GPT actually means generative pre-trained transformer. A transformer is a type of artificial neural network that is trained using deep learning, a term that refers to the many layers (deep) in neural networks. Deep learning has contributed to many recent advances in artificial intelligence.

Nevertheless, some characteristics distinguish basic models from previous generations of deep learning models. To begin with, they can be trained on extremely large and diverse sets of unstructured data. For example, a type of basic model called a large language model can be trained on a huge amount of text that is publicly available on the Internet and covers many different topics. While other deep learning models can work on significant amounts of unstructured data, they are usually trained on a more specific set of data. For example, a model might be trained on a particular set of images so that it can recognize certain objects in photos.

In fact, other deep learning models can often perform only one such task. They may, for example, either classify objects in a photo or perform another function, such as making predictions. In contrast, a single Foundation model can perform both of these functions and generate content. Foundation models combine these capabilities by learning patterns and relationships from the broad learning data they ingest, which, for example, allows them to predict the next word in a sentence. Here’s how ChatGPT can answer questions about different topics and how DALL – E 2 and Stable Diffusion can generate images based on descriptions.

Given the versatility of the base model, companies can use the same model to implement multiple business use cases, which is rarely achieved with earlier deep learning models. The base model, which includes information about a company’s products, can potentially be used both to answer customer questions ‘ and to support engineers in developing updated versions of products. As a result, companies can maintain applications and realize their benefits much faster.

However, because of the way modern base models work, they are not suitable for all applications. For example, large language models can be prone to hallucinate “, ” or answer questions with plausible but incorrect statements (see Sidebar below ” Using generative AI responsibly “). In addition, basic reasoning or answer sources are not always provided. This means that companies should be cautious when integrating generative AI without human control in applications where errors can cause harm or where explainability is needed. Generative AI is also not currently suitable for directly analyzing large amounts of tabular data or solving complex numerical optimization problems. Researchers are working hard to address these limitations.

The use of generative AI responsibly

Key Ideas

The impact of generative AI on productivity could add trillions of dollars to the value of the global economy. Our latest research shows that generative AI could add the equivalent of $2.6 trillion to $4.4 trillion annually in the 63 use cases we analyzed – for comparison, the entire United Kingdom’s GDP in 2021 was $3.1 trillion. This would increase the impact of all artificial intelligence by 15-40 percent. This estimate would roughly double if we included the impact of implementing generative AI in software that is currently used for tasks other than those in use.

About 75 percent of the value that generative AI options can provide comes from four areas: customer management, marketing and sales, software development and research and development. In 16 Business Functions, we looked at 63 use cases in which technology can solve specific business problems in ways that produce one or more measurable results. Examples include the ability of generative AI to support customer interactions, generate creative content for marketing and sales, and compose computer code based on natural language prompts, among many other tasks.

Generative AI will have a significant impact on all industries. Banking, high-tech and life sciences are among the industries that could see the greatest impact as a percentage of their revenues from generative AI. For example, across the banking industry, the technology could generate value equal to an additional $200 billion to $340 billion per year if the use cases are fully realized. In retail and consumer goods, the potential impact is also significant at $400 billion to $660 billion per year.

Generative AI can change the anatomy of work by empowering individual workers by automating some of their individual actions. Today’s generative AI and other technologies can automate jobs that now consume 60 to 70 percent of employees. In contrast, we previously estimated that technology could automate half of the time employees spend at work.4The acceleration in the potential for technical automation is largely due to the increased ability of generative AI to understand natural language, which is required for work activities that account for 25 percent of total work time. Thus, generative AI has a greater impact on knowledge work associated with occupations that have higher wages and educational requirements than other types of work.

The pace of workforce transformation is likely to accelerate, given the increasing potential for technical automation.  Our updated deployment scenarios, including technology development, economic feasibility, and distribution timelines, estimate that half of today’s work activities could be automated between 2030 and 2060, with a midpoint in 2045, or about a decade earlier than our previous estimates.

Generative AI can significantly increase productivity in the economy, but it will require investment to support workers as they change careers or change jobs. Generative AI can provide 0.1 to 0.6 percent productivity growth annually through 2040, depending on the pace of technology adoption and reallocation of work time to other activities. Combining generative AI with all other technologies, work automation could add 0.2 to 3.3 percentage points annually to productivity growth. Nevertheless, workers will need support in learning new skills, and some will change occupations. If worker transitions and other risks can be managed, generative AI can contribute significantly to economic growth and support a more sustainable, inclusive world.

The era of generative AI is just beginning. The excitement about this technology is palpable, and the early pilots are compelling. But fully realizing the benefits of the technology will take time, and business and community leaders still have significant challenges to address. These include managing the risks inherent in generative AI, determining what new skills and capabilities the workforce will need, and rethinking core business processes such as retraining and developing new skills.

Where the value of business lies

Generative AI – is a gradual change in the evolution of artificial intelligence. As companies seek to adapt and adopt it, understanding the technology’s potential to create value for the economy and society as a whole will help make critical decisions. We used two additional lenses to determine where generative AI with its current capabilities could provide the most value and how big that value might be (Application. 1 ).

                                          Application. 1

First, Lens scans use cases for generative AI that organizations can adopt. We define a ” use case ” as a targeted application of generative AI to a specific business task that results in one or more measurable outcomes. For example, an example of a marketing use case is the application of generative AI to create creative content such as personalized emails, measurable results that potentially include reduced costs of creating such content and increased revenue from improved effectiveness of higher quality content at scale. We identified 63 generative AI use cases covering 16 business functions that could provide total value in the range of $2.6 trillion to $4.4 trillion in economic benefits annually when applied across industries.

This would add 15 to 40 percent to the $11 trillion to $17.7 trillion economic value that we estimate can now be unlocked.

Traditional advanced analytics algorithms

and machine learning are very effective at performing numerical and optimization tasks, such as predictive modeling, and continue to find new applications in a wide range of industries. However, as generative AI continues to evolve and develop, it can open up entirely new frontiers in creativity and innovation.

In addition to the potential value generative AI can provide in use cases for specific functions, the technology can add value to the entire organization by revolutionizing internal knowledge management systems.

Impressive mastery of generative AI  natural language processing can help employees gain stored internal knowledge by formulating queries in the same way that they can ask a person a question and engage in an ongoing dialogue. This can enable teams to access relevant information quickly, allowing them to make better-informed decisions quickly and develop effective strategies.

What is the difference between machine learning and artificial intelligence?

Artificial Intelligence – That’s pretty much what it’s like-the practice of getting machines to mimic human intelligence to perform tasks. You’ve probably interacted with AI even if you don’t understand it — voice assistants like Siri and Alexa are based on AI technology, as are customer service chatbots that help you navigate Web sites.

Machine Learning – is a type of artificial intelligence. Through machine learning, practitioners develop artificial intelligence with models that can “learn from” data models without human direction. The unmanageably huge volume and complexity of data (, unmanageable by humans anyway ) that is now being generated has increased the potential of machine learning as well as the need for it.

What are the main types of machine learning models?

Machine learning is based on a number of building blocks, starting with classical statistical methods developed between the 18th and 20th centuries for small data sets. In the 1930s and 1940s, computing pioneers – including theoretical mathematician Alan Turing – began working on basic machine learning methods. But these methods were confined to laboratories until the late 1970s, when scientists first developed computers powerful enough to mount them.

Until recently, machine learning was largely limited to predictive models used to observe and classify patterns by content. For example, a classic machine learning problem is to start with an image or several images of, say, adorable cats. The program would then identify patterns among the images and then carefully learn random images for images that would match the pattern of the adorable cat. Generative AI was a breakthrough. Instead of simply perceiving and categorizing a photo of a cat, machine learning can now create an image or text description of the cat on demand.

How do text-based machine learning models work? How are they trained?

The first machine learning models for word processing were trained by humans to classify different inputs according to labels set by researchers. One example would be a model trained to label social media posts as positive or negative. This type of learning is known as supervised learning because a human is responsible for “teaching” the model what to do.

Next-generation text-based machine learning models are based on what is known as supervised learning. This type of learning involves feeding the model a huge amount of text so that it can generate predictions. For example, some models can predict, based on a few words, how a sentence will end. With the right amount of sample text-say, a wide range of the Internet-these text models become quite accurate. We see how accurate with the success of tools such as ChatGPT.

What does it take to build a generative AI model?

Creating a generative AI model has, for the most part, been a serious undertaking to the extent that only a few high-tech heavyweights have made the attempt. OpenAI, the company behind ChatGPT, the former GPT and DALL-E models, has billions in funding from big-name donors. DeepMind is a subsidiary of Alphabet, Google’s parent company, and Meta has released its Make-A-Video product based on generative AI. These companies employ some of the best computer scientists and engineers in the world.

But it’s not just talent. When you ask a model to train across almost the entire Internet, it’s going to cost you. OpenAI has not published exact costs, but estimates show that GPT-3 trained about 45 terabytes of text data — that’s about a million feet of bookshelf or a quarter of the entire Library of Congress — at an estimated cost of several million dollars. These are not resources your garden-variety startup can access.

What kinds of products can a generative AI model produce?

As you may have noticed above, output from generative AI models can be indistinguishable from human content, or they may seem a bit odd. The results depend on the quality of the model-as we have seen, ChatGPT’s results so far surpass those of its predecessors-and the fit between the model and the use case, or input.

ChatGPT can produce what one commentator called a “solid” essay comparing the theories of nationalism from Benedict Anderson and Ernest Gellner–in ten seconds. It also produced an already famous passage describing how to remove a peanut butter sandwich from a VCR in the style of the King James Bible. AI-generated art models like DALL-E (, called a mixture of the surrealist Salvador Dali and the cute Pixar robot WALL-E ), which can create strange, beautiful images on demand, like Raphael painting Madonna and a baby eating pizza. Other generative AI models can create code, video, audio or business simulations.

But the outputs aren’t always accurate – or relevant. When Priya Krishna asked DALL-E 2 to come up with an image for Thanksgiving dinner, it produced a scene where a turkey was decorated with whole limes, set next to a bowl of what appeared to be guacamole. For its part, ChatGPT seems to be having trouble counting or solving basic algebra problems-or, indeed, overcoming the sexist and racist prejudices that lurk in the streams of the Internet and society more broadly.

What problems can a generative AI model solve?

You’ve probably seen these generative AI tools (toys? ) like ChatGPT can generate endless hours of entertainment. The opportunity is clear for business as well. Generative AI tools can create a wide range of credible emails in seconds, and then respond to criticism to make the writing more fit for purpose. This has implications for a wide range of industries, from IT and software organizations that can benefit from instant, mostly correct code generated by AI models to organizations needing marketing copy. In short, any organization that needs to produce clear written materials could potentially benefit. Organizations can also use generative AI to create more technical materials, such as higher-resolution versions of medical images. And with the time and resources saved here, organizations can leverage new business opportunities and the ability to create more value.

We’ve seen that developing a generative AI model is so resource intensive that it’s out of the question for all but the largest and most affluent companies. Companies that want to use generative AI for work have the option of either using generative AI out of the box or fine-tuning them to perform a specific task. For example, if you need to prepare slides according to a particular style, you can ask the model to ” learn ” how headlines are typically written based on the data on the slides, then feed the slide data and ask them to write the appropriate headlines.

Considerations for everyone

As artificial intelligence (AI) continues to evolve and be applied in various fields, it has the potential to significantly affect the course of history. Here are some of the possible changes that AI can bring:

  1. Workforce Transformation: The introduction of AI can lead to the automation of routine tasks and processes, which in turn can change the structure of the workforce. Some jobs may be replaced by AI, while other areas may require new skills and specializations.

  2. Improved performance and efficiency: AI has the ability to analyze and process large amounts of data at high speed and accuracy. This can lead to improved productivity and efficiency in a variety of industries, including manufacturing, healthcare, finance, and others.

  3. A revolution in medicine and health care: AI can play a significant role in the fields of medicine and health care, such as in disease diagnosis, individualized treatment, and the development of new drugs. This could lead to more accurate diagnoses, improved health care, and increased life expectancy.

  4. The Transport Revolution: AI can play an important role in the development of autonomous vehicles and optimization of transportation systems. It can affect road safety, reduce travel time, and improve transport efficiency.

  5. Changing social interactions:AI can also affect our social interactions and society as a whole. For example, the development of robots and virtual assistants can change the way we interact with technology and with each other.

History has shown that new technologies can change society. Artificial intelligence has already changed the way we live and work!

But technology can also pose new and serious challenges. Stakeholders must act–and quickly, given the pace at which generative AI can be adopted–to prepare to address both the opportunities and the risks.

Using generative AI - considerations

Bottom line, We hope this research has contributed to a better understanding of Generative AI’s ability to add value to company operations and drive economic growth and prosperity, as well as its potential to dramatically transform the way we work and our purpose in society. Companies, politicians, consumers and citizens can work together to ensure that generative AI delivers on its promise to create significant value while limiting its potential to destroy lives and livelihoods. The time to act is now.

Yet generative AI represents another promising leap forward and a world of new possibilities.


Konovalov Alexey – partner at Marrbery, where Natalya Shevchenko – consultant; Marina Krivosheya – senior expert; and Maria Zankovetskaya – consultant.

The authors would like to thank the following individuals:

Galina Kovalenko, Kirill Snitko, Vasily Litvinenko, Anna Grigorenko, Alexander Melnik, Olesa Savchenko, Oleg Koval, Denis Gordienko, Bogdan Timchuk, Lilia Kravka

and others who helped in the process.

Want to know more about AI?

Keep up to date with our latest statistics