Generative AI is fueling hopes of a new economic boom. Many observers fail to grasp the fairest way of how AI is likely to generate value, and they cling to frames of thought they are familiar with. Thus, Goldman Sachs combines the rise of artificial intelligence with disciplined corporate governance characterized by a dramatic reduction in labor costs and the centralization of decision-making processes. We believe this approach is irrelevant and potentially even dangerous.
On March 26, the Economic Research Laboratory Goldman Sachs He shared his insights into the potential consequences of the emergence of generative AI on economic activity. Multiplied by the fantastic advertisements, this report was taken up extensively in the media. According to Goldman Sachs, “ We are on the cusp of a rapid acceleration in the automation of tasks that will reduce labor costs and increase productivity This assertion does not allow us to take a measure of what lies ahead, we allow ourselves to quote triple figures. The authors of the report claim that artificial intelligence will lead to 300 million jobs lost in the world and in 7% annual increase in GDP. In 2030, it will represent US investments in artificial intelligence 1% of the GDP, i.e. one-third of the country’s spending on research and development. it’s a lot!
We believe that the financial ideology behind this report significantly biases the stated expectations. First of all, since labor costs are a very important cost component in many companies, task automation can emerge as an important lever for profitability. It often seems easier to cut costs than to expand markets. Moreover, the growing tendency of “scalable” business models encourages the search for growth that will not be accompanied by hiring. Finally, in all developed countries the effective rates of taxes on labor are much higher than those applied to fixtures and equipment. So it’s not surprising that a large financial institution like Goldman Sachs sees generative AI only as a way to automate work and the main focus of corporate investment for years to come.
However, this ideology overlooks a certain number of economic and entrepreneurial realities. It seems important to us to remember this.
First, growth gross domestic product It is not independent of its distribution. If AI replicates and automates existing human capabilities, machines become better substitutes for human labor and workers lose their bargaining power for the wealth created. A fully automated economy could, in principle, be structured in such a way as to widely redistribute production profits, even those no longer strictly necessary for value creation. However, the beneficiaries would be in a weak position to prevent a change in distribution that would leave them with little or nothing. They will depend precariously on the decisions of those who control the technology. This will open the door to an increased concentration of wealth and power. in 1941US Supreme Court Justice Louis Brandeis said: We have to choose. Democracy or wealth can be concentrated in the hands of a few, but we cannot have both. Fortunately, two centuries of technological progress have increased the economic value of the work of all categories of employees. The economic value of an hour of work for a forklift driver is greater than the economic value of an hour of work for a forklift driver. An administrative agent who masters the set of software used by his company receives a wage Much higher than the wages of former employees in the writings.The beneficial effects of productivity gains resulting from technology have therefore spread widely among the population.The economic value of an hour of human labor, measured by average wages, has increased more than tenfold in two centuries.The increase in Human capabilities that technology has enabled have been an engine of economic growth for two centuries.On the contrary, when a person loses the opportunity to earn income from his work, the costs go beyond the newly unemployed and affect many other members of their community and society at large.In this sense, The automation of labor brings less growth than the increase in human capabilities through technology.Wide distribution of the fruits of technological progress supports economic growth.This shared prosperity also ensures popular approval of the political institutions of our democracies.
Then, automating repetitive or formalizable tasks rarely results in job losses. In 2018, Brynjolfsson, Mitchell, and Rock analyzed the impact of introducing machine learning on 2069 work activities, 18156 tasks and 964 occupations defined in the O*NET database. The results of their study are justified. (1) Most professions, regardless of the sector under consideration, have tasks that can be automated thanks to machine learning. (2) No profession can be fully automated. (3) Exploiting the potential of machine learning requires a major shift in job content, and the invention of new organizational structures. Some indications are that the value creation authorized by AI stems, at least in part, from the importance of new organizations resulting from it. imagine that Jeff Bezos Existing libraries are “automated” simply by replacing cashiers with robots. This may have cut costs a bit, but the overall effect would have been limited. Instead, Amazon has reinvented the concept of a bookstore by bringing together humans and machines in a way it has never been before. As a result, Amazon offers a much wider selection of products, ratings, reviews, and advice. Retail access is allowed 24 hours a day, 7 days a week, from the comfort of customers’ homes. The power of technology is not to automate the work of humans within the existing retail bookstore concept, but to reinvent and augment the way customers find, rate, purchase and receive books, and later other retail products.
Finally, until now, no human brain can contain even a small part of the knowledge needed to run a medium-sized company, so decisions are distributed and decentralized. Distribution of decisions and decentralization is the source of a series of innovations related to products and internal company operations. Rooted in the tangible experience of individuals, these innovations are difficult to imitate by competitors who neither understand their root nor their purpose. On a day to day basis, these are the company’s main source of differentiation and eliminating any price encounter risks. Moreover, organizations theory has already shown that modular organizations made up of discrete entities that have large discretionary spaces and act independently within the same whole constitute the best response to uncertainty. On the other hand, the devastating effects of wrong decisions taken at the top of the central organizations are systemic in nature and are likely to send the company to the bottom. And no one ever claimed that generative AI was omniscient. Therefore, we believe that organizations of the future will have to reconcile the massive use of generative AI with the decentralization of decision-making processes. In this case, they will find their means of differentiation and the ability to avoid the dangers of an uncertain environment.
Associating technological progress with the complete automation of human labor and the elimination of jobs is a historical misconception. Since the beginning of the industrial revolution, technology has increased human labor capabilities and changed job content. The resulting increases in productivity have added economic value to the work of millions of employees, and this has supported the economic growth of major democracies.
Generative AI is part of this line. Unable to cover all the tasks included in a job, generative AI represents a new stage in the augmentation of human labor by means of technology. In the coming years, it will contribute to the redefinition of job content and will strongly influence the organization of companies. In this context, the ability of AI to generate value will depend on the ability of managers to identify effective organizations that are hard to imitate and that are equipped to deal with uncertainty.
 Brynjolfsson, E., Mitchell, T., & Rock, D. (2018, May). What can machines learn and what does it mean for professions and the economy? In AEA Papers and Procedures (Vol. 108, pp. 43-47).
Tribune by Eric Brown – Associate Professor – BA INSEEC and Pascal Montignon – Research Chair Director for Digital, Data Science and Artificial Intelligence – OMNES EDUCATION
<< اقرأ أيضًا: TRIBUNE | ما الذي يجلبه ChatGPT والذكاء الاصطناعي التوليفي حقًا إلى الذكاء الاصطناعي التحليلي؟ >>>