Generative AI could boost productivity by $4.4 trillion annually

30 August 2023 4 min. read

While the tangible results of AI are hard to come by at present, the hype surrounding the technology is going from strength to strength. According to a new study from McKinsey & Company, generative AI could boost global productivity by more than $4 trillion every year.

Responses to a simple web search for precisely how much AI is adding to the global economy remain vague. When asked “how much is AI contributing to the global economy now”, the top results will invariably relate to the latest hypothesis of what the technology ‘could’ be worth, rather than what it actually ‘is’.

After some scrolling, searchers may unearth estimates from GrandViewResearch suggesting the global AI market size around $93.5 billion in 2021, and if the predicted compound annual growth rate of 38.1% from 2022 until 2030 is accurate, that should have risen to a value of $136.6 billion by now. But it is admittedly far less eye-catching than the latest study from McKinsey & Company, touting the mouthwatering prospect to businesses of realising an extra $4.4 trillion in productivity annually.

Generative AI could boost productivity by $4.4 trillion annually

In particular, that figure relates to the potential of generative AI. Proponents of the technology suggest that generative AI enables users to quickly generate ‘new’ content – though originators of the art, text and music which the AI is ‘trained’ with in order to churn out variants of dispute this – based on a variety of inputs. Inputs and outputs to these models can include text, images, sounds, animation, 3D models, or other types of data.

On the basis of its predictions, McKinsey suggests that by the end of the decade, generative AI could roughly be creating between $2.5 trillion and $4.4 trillion annually – the equivalent value of one whole UK economy, give or take the odd billion. The researchers determined this by identifying 63 generative AI use cases spanning 16 business functions, which the firm suggested could deliver those economic benefits annually when applied across all industries.

In line with much of the other research around generative AI, the strategy consultant suggests that the technology boasts the greatest opportunities when relating to creative functions. According to McKinsey, using generative AI in the sales segment – helping to generate emails to drum up new business, among other uses – could generate close to a $500 billion impact, with less than 10% of all functional spend. Similarly, the marketing function – generating campaigns, images, and text content – could yield more than $450 billion in impact, against 10% of functional spending.

Generative AI could boost productivity by $4.4 trillion annually

"The real impact could be even greater if the technology was integrated into software such as word processing programs or chatbots, which would free up working time for other tasks. The development of generative AI opens a new era of technological innovation – it will be a tool for increasing productivity and accelerating global economic growth," said Anna Katariina Wisakanto, a digital consultant and generative AI expert at McKinsey.

To that end, McKinsey modelled a further set of scenarios to see how the current capabilities of AI could also cause incremental, non-direct benefits that would end up boosting the world’s economic performance. Estimating when generative AI could perform each of more than 2,100 “detailed work activities”, including “communicating with others about operational plans or activities”, McKinsey asserted that while some generative AI impacts could overlap with other areas of improved labour productivity, netting out this overlap, the total economic benefits of generative AI when the technology is applied across knowledge workers’ activities amounts to $6.1 trillion to $7.9 trillion annually.

Still, not everyone is convinced of the idea that generative AI will inherently improve from now until 2030. Some experts warn that the technology’s use of the internet to source training data is already producing signs of ‘model collapse’ – as generative AI continues to be trained by misinformation – a growing portion of which generative AI itself is already producing. If the huge predictions currently being thrown about regarding generative AI’s ability to save time and ramp up productivity, concerns like these will need to be addressed. Otherwise, the value of generative AI might end up eternally being a “could be”, rather than an “is”.