McKinsey releases generative AI reports predicting human levels by 2030

Source: Shin Zhiyuan

McKinsey AI report released, generative AI has made rapid progress and huge economic benefits, and the future cannot be underestimated.

McKinsey Blockbuster Report Released!

The core conclusion is one sentence: AI will reach human levels sooner than thought, and the median projection is before 2030.

You know, compared with people's predictions in 2017, the new report highlights an optimism.

The figure above is the final result of the report, which we will talk about one by one later.

Report Roundup

At the outset, the report begins with a perfect summary of how much technology has affected our lives today.

In short, AI has long penetrated into all aspects of our lives.

When DeepMind came up with AlphaGo in 2016 and defeated world champion Lee Sedol, AI once entered our field of vision overwhelmingly, but because it was only limited to the game of Go, it slowly faded out as soon as the limelight passed.

But this year is different.

Not to mention ChatGPT, which has exceeded the sky in terms of users, only generative AI products such as Copilot, Stable Diffusion and so on, have swept our lives like a storm.

What's different this time is that these AI tools are available to everyone. Everyone can create with ChatGPT, use Midjourney to make graphics, and use Copilot to do PPT.

ChatGPT equipped with GPT-4, all performance directly takes off from GPT-3.5. Then there's Anthropic's Claude, which can process 100,000 tokens a minute (about the length of a novel), and Claude's generation in March this year is almost one-tenth the current performance.

The report focuses on the speed at which AI is developing, rising in just a few months.

In this report, generative AI is defined as an application built with a base model. The base model has a large number of new features in images, video, audio, code, etc., and the performance of the original functions has also been greatly improved.

According to the report, our understanding of the magnitude of generative AI's capabilities is still in its infancy.

That's why McKinsey made a report in order to better understand the future of generative AI.

Economic and Social Impact

Companies are experimenting with generative AI to quickly adapt their workflows to new technologies.

The report points out that it is necessary to have a thorough understanding of what generative AI will bring to our overall social and economic development.

In the chart below, the report uses two complementary perspectives to determine where and how much value is currently being brought to the capabilities of generative AI.

Lens 1 in the image above is a large scan of businesses that can use generative AI. This is called a "use case".

Let's say a use case in marketing is: Generative AI is applied to generate content such as personalized emails, measuring results such as reducing the cost of generating such content and increasing revenue by increasing the effectiveness of high-quality content at scale.

As a result, the report identifies a total of 63 generative AI use cases covering 16 business functions that could bring between $2.6 trillion and $4.4 trillion annually in economic benefits if applied across industries.

Look at a lot.

That's a 15 to 40 percent increase from the current estimated economic value of $11 trillion to $17.7 trillion. The latter is McKinsey's forecast for 2017.

Lens 2 complements Lens 1, with the report analyzing the potential impact of generative AI on some 850 occupations.

Experts simulated a variety of scenarios to estimate when generative AI would be able to perform each of the more than 2,100 jobs that make up the global economy — which could include tasks such as communicating with others about operational plans or activities.

In this way, we can estimate how generative AI will affect labor productivity for all the jobs currently performed by all the global workforce, with existing capabilities.

Some of these impacts overlap with the cost reductions mentioned in Lens 1, and the report therefore assumes that the cost reductions are the result of increased labor productivity.

Excluding this overlap, the total economics of generative AI amount to $6.1 trillion to $7.9 trillion per year, as shown in the chart below.

Future Potential

While the economic benefits are already considerable, the report says they go far beyond that.

Let's talk about potential.

Generative AI is likely to have an impact on the functions of most businesses. However, if we measure the technical impact as a percentage of the cost of the function, a few functions stand out, as shown in the chart below.

McKinsey analyzed 16 business functions and found that only four functions—customer operations, marketing and sales, software engineering, and research and development—accounted for about 75 percent of the total annual value of generative AI use cases.

Simply put, from a technical perspective of the work itself, not all businesses benefit from AI to a large extent.

The report notes that the potential value of generative AI was much lower in several areas of work, including manufacturing and supply chain, when evaluating various use cases for AI.

The main reason is due to the nature of generative AI itself.

In addition to the potential value that generative AI can bring in specific use cases, generative AI can also bring value to the entire company by revolutionizing knowledge management systems within the enterprise.

We all know that generative AI has strong natural language processing capabilities and can help employees more easily query and retrieve internal knowledge stored by the company.

Clearly, this enhances the team's ability to quickly access relevant information, enabling them to quickly make more informed decisions and develop effective strategies.

Before the advent of generative AI, the same work may have taken workers a whole day to do, and generative AI must have produced huge benefits after undertaking these tasks.

In addition, generative AI can increase value by working with workers, accelerating their productivity, and enhancing their ability to do their jobs.

I don't say whose DNA moved, even this article was generated by the editor with AI (no).

Of the 63 use cases analyzed in the report, generative AI has the potential to create a total value of $2.6 trillion to $4.4 trillion across industries.

Of course, the exact impact depends on a variety of factors, such as the combination of different functions, the importance of each, and more importantly - the revenue scale of the industry itself, as shown in the figure below.

For example, generative AI can reportedly bring about $310 billion in additional value to the retail industry, including car dealerships, by improving capabilities such as marketing and customer operations.

In contrast, much of the potential value in high-tech comes from generative AI's ability to improve the speed and efficiency of software development, as shown in the chart below.

The report estimates that this number will become increasingly spectacular in the future — as AI's capabilities advance by leaps and bounds.

Since 2017, the McKinsey Global Institute has been analyzing the impact of automation on different work activities, and they have also modeled various scenarios for the adoption of technology.

At the time, they estimated that workers spent at least half of their time adapting and automating existing technologies, what we call the automation potential of technology.

Experts also simulated a range of possible scenarios to determine the rate of adoption of these technologies and their impact on work activities in the global economy.

First, the large-scale adoption of technology will not happen overnight. It takes time for technology in the lab to translate into automation of specific work activities.

At the same time, if the cost of automation is higher than the cost of labor, it is obviously not feasible.

Finally, even if it does, it will take time to roll it out on a larger scale.

That's what the report focuses on. How much potential generative AI has for automation in production and life, and how much work efficiency can be improved.

The report predicts that, based on the current performance of generative AI, its capabilities in all aspects will reach human performance faster than previously estimated, as shown in the chart below.

The institute previously thought that 2027 was the earliest year that technology could reach the intermediate level of human natural language understanding, but in the latest report, this time has been brought forward to 2023.

Theoretically, by integrating existing technologies, the total proportion of automation at this stage has increased from about 50% to 60-70%.

Moreover, due to the rapid development of generative AI natural language capabilities, the curve of technological development potential is quite steep.

The chart below shows the 2017 forecast and the latest forecast, and from the curve we can easily see how the word "optimistic" is written.

Latest forecasts

Forecast for 2017

The chart below shows how much the activities that workers perform on a daily basis will change in the report, with the latest forecast at the top and the forecast for 2017 below.

Latest forecasts

Forecast for 2017

Experts predict that generative AI is likely to have the greatest impact on knowledge work, especially activities involving decision-making and collaboration, which previously had the lowest potential for automation, as shown in the chart below.

The report estimates that the potential for automation of expertise jumped by 34 percentage points, while the potential for automation of management and talent development rose from 16% in 2017 to 49% in 2023.

In addition, generative AI is also able to understand natural language and use it for a variety of activities and tasks, which goes a long way towards explaining why automation has so much potential.

In the economic sphere, about 40 percent of the activities engaged in by workers need to reach at least the median level of human understanding of natural language.

As a result, many work activities involving communication, supervision, recording, and human interaction have the potential to be automated through generative AI, accelerating job transformation in occupations such as education and technology, where automation potential was previously expected to emerge later, as shown in the chart below.

In addition to the above, the McKinsey report also analyzes from other dimensions.

Due to space limitations, I will not list them all.

**Where to go from here? **

The above analysis can be said to focus on the overall appearance of the industry.

To reflect the grounding of the report, the final section is the impact of generative AI on individuals and how each of us should face it.

As new technologies evolve, stakeholders must act to prepare for opportunities and risks, the report says.

The main concern is also a cliché, such as hallucination problems, intellectual copyright issues of data used in training, and so on.

The report predicts that at least a quarter to a third of jobs will change over the next decade under the median projection. For different roles in different people, we have to respond very differently.

For companies and business leaders, how can they harness the potential value of generative AI while managing the risks it poses.

How will generative AI and other AI technologies change the occupations and skill sets required of a company's workforce in the coming years? How will companies implement these shifts in hiring programs, restraining programs, and other aspects of HR?

Can companies play a role in ensuring that technology is not used in negative ways that could harm society?

How can companies share their experiences in promoting the use of generative AI within and across industries with governments and society in a transparent way?

These are questions that managers need to explore.

What does generative AI mean for future workforce planning for decision makers in government?

How can workers be provided with the necessary policy support when their activities change over time?

Can new policies be formulated or existing ones revised to make AI more socially valuable?

Finally, as every worker, consumer, and citizen, how should we pay attention to the development of new technologies? Where do we get the right and fair information?

How can individuals balance the convenience and impact of generative AI?

How do we, as individuals, express our demands in the decision-making process?

Many issues urgently require our in-depth consideration.

In short, this report provides a comprehensive look at the significant impact of the generative AI explosion on our society, especially on the economy.

Resources: