Chat GPT is causing a lot of emotions today. Not only because it has surpassed the level of 100 million active users in record time – two months – (for comparison, TikTok needed nine months for that). But also because of the concerns it has raised. These are reflected in appeals to temporarily halt the development of AI-based technologies.
Lack of guidelines
I have already referred several times to the risks associated with technology, especially the issue of not keeping up with regulatory frameworks and the unpredictability of the effects of using certain solutions. Although I am aware that implementing this practically will not be easy, I fully agree with the demand that we should give ourselves time to create legal frameworks for the further development of artificial intelligence. Even if they do not have a global character – I presume that China, which widely uses these technologies to monitor its citizens, will not want to sign them – it is still worth using this temporary “window of interest” to create at least the rudiments of rules for the use of AI, and to establish some sort of guidelines.
At the same time, we must be aware that even if the development of certain AI-based solutions is stopped for a while, the use of artificial intelligence itself will quickly become widespread, and for many reasons: it is already proving useful in many areas, its development can bring tangible benefits to us humans (e.g., in medicine), and the main driving force of the world remains “greed” (and you can make a lot of money from AI). Therefore, in addition to ethical issues, we must also look at other aspects of AI usage, including those related to economic and social processes. This is precisely what this text is devoted to.
The power of general purpose technology
The acronym GPT stands for Generative Pre-trained Transformer. However, GPT is also an abbreviation for another important concept in the field of technology: General Purpose Technology, which refers to technologies with general applications. It is commonly accepted that AI-based solutions, including GPT Chat, represent the next generation of such technologies. Under this unassuming term lies an enormous power to transform societies and economies. This is evidenced by the transformations that have taken place under the influence of other GPTs, such as electrification or computerization. The huge impact of these technologies is due to their universal application and the fact that they translate into the generation of complementary innovations, resulting in rapid and profound transformations affecting vast areas of human life and activity.
Issues related to general purpose technologies cover a very wide range of topics. In this post, I would like to focus only on the aspect of their impact on the macroeconomic sphere. I will attempt to answer the question of whether a new general-purpose technology will translate into a higher quality of life for us.
Among the main channels of impact on the economy – in particular the size and structure of GDP – two approaches are usually considered: supply-side and demand-side. The former takes into account the impact of AI on production processes; AI-based solutions will lead to the reallocation of production factors (capital, labor) from one activity to another. They should also lead to the more efficient use of resources (increased productivity), and thus allow for the generation of higher GDP.
In the demand-side approach, the impact of AI is considered in terms of:
- International flows – on the one hand, on the structure of these flows (increased importance of data and related or based-on data services and the accompanying relative decline in the importance of goods flows), and on the other hand, on the directions of individual streams. The latter changes cannot be considered in isolation from other global processes, such as geopolitics or regionalization. It is worth noting that although the global impact of this element remains neutral, it is of great importance from the point of view of the redistribution of benefits resulting from the use of AI on individual countries.
- Consumption – it may seem that the use of AI-based technologies should translate into higher consumption. However, this is not so certain. Much will depend on income distribution. In recent decades, the increasingly intensive use of technology has led in many countries to a decrease in the share of wages in total income. To compensate for the negative effects of this process on consumption in some countries, credit has become an alternative. However, today the level of household debt is high, especially in developed countries, so it is difficult to expect that it will be able to play a similar role as in the past in stimulating consumer demand. Therefore, institutional solutions that provide households with the possibility of obtaining an appropriate level of income will be crucial for the final effects (more on this below).
- Investments – it is expected that, as was the case with electrification and computerization, AI-based solutions will stimulate new streams of investment, related not only to AI itself (its development and dissemination), but also to the need to replace networks, devices, and other elements of production processes so that they can operate in systems based on/using AI. There is also an expectation of a lasting increase in investment in human capital, new skills, and competencies.
Summa summarum, estimates from various consulting firms* indicate that, at the net level – taking into account the effects of cannibalization of existing markets of goods and services – artificial intelligence will allow for the generation of additional GDP (income). However, the distribution of benefits remains a completely separate issue, for at least four reasons.
GPT does not necessarily mean better quality of life for everyone.
The first issue relates to the accessibility of benefits resulting from the use of AI in various services, such as healthcare or education. It may turn out that advanced solutions in the field of diagnosis or therapy selection (healthcare) or personalized learning (education) will not become part of the universal (public) offer, but only paid services for those who can afford it.
The second issue is related to the ability to match qualifications in a given society to the needs of an AI-based economy. It is virtually certain that this will be achieved to varying degrees in different countries. Countries where this is not achieved will lose good (well-paid) jobs to others.
The third issue is related to the fact that AI-based solutions may also have adverse effects on the quality of our lives. If we do not create appropriate safeguards, we may become powerless in the face of permanent surveillance. We may also have to live in fear of cyber threats.
The fourth and final issue is related to the redistribution of income generated in connection with the use of AI-based solutions. Assuming that institutional solutions such as those in recent decades will still be applied, it is highly likely that mass use of AI will result in widening income inequalities. This is due to the fact that in income terms, the biggest beneficiaries of using AI will be a relatively narrow group of owners of leading technologies. Why? Because by its nature – if not restricted by certain limiting solutions – the market for AI-based solutions tends to operate based on two key principles:
- “the winner-takes-all”, which translates into the functioning of digital services in the form of monopolies or oligopolies,
- cannibalization of traditional (“analog”) business models, which by their nature had a more competitive character and provided a wider range of jobs.
Technology and Institutions
That is why it is so important to learn from the experiences that have emerged from previous revolutions that accompanied the dissemination of older general-purpose technologies. We need to redesign the socio-economic system so that the benefits of the latest technological solutions are widely spread throughout society, both in terms of access to certain goods/services and income. Also, in terms of security against abuses related to the misuse of AI-based technologies.
To make this happen, we need a variety of decisions regarding mechanisms such as:
- securing data from unauthorized use,
- valuation and distribution of benefits from data – we are their main source, but today, tech companies are mainly their beneficiaries. Not to mention the abuses that accompany this.
- pricing public goods (e.g., basic research, R&D expenditures) used by private companies to build their business models based on AI
- strengthening the primary distribution of income in favor of households, including by providing a wider range of income sources (e.g., income from prosumer models in the energy sector, various forms of shareholding, etc.)
- secondary income redistribution through the budget (this largely concerns taxation issues, including taxation of technology, but also social solutions)
- pricing work and defining the concept of work itself (perhaps we need to consider pricing and remunerating not only salaried work but also other activities, such as domestic or social work).
In summary, as I have repeatedly written, in the case of smart societies, adjustments in institutional solutions are the proper response to technological progress and the resulting social transformations. If we want the revolution associated with artificial intelligence to be beneficial for us as a society, we must start designing new solutions today. These should address both the direction of technological progress and the economic and social consequences of its use.
The ability to rebuild the socio-economic system in the face of technological revolution is – as history confirms – key to success and experiencing global progress. The lack of such an ability leads a country to marginalization. With the current level of technological advancement, the stakes are even higher, as the survival of our civilization is at risk. That is why the dimension of international cooperation and coordination has become so important.
* Estimates of this type can be found, among others, in:
– Accenture – „How AI boosts industry profits and innovation”,
– Analysis Group – “Global Economic Impacts Associated with AI”,
– McKinsey – “Notes from the AI frontier: modelling the impact of AI on the World Economy”,
– PWC – “The macroeconomic impact of AI”
Leave a Reply