Human activity is highly dependent on information. What to eat? Which movie to watch? Where to travel? What to study? Where to work? It all depends on information. The same applies to economic activity, where information has a crucial influence on business strategy: which products and services should we provide? Where should we produce them? How should we sell them? What is the marketing strategy? At which price? Et cetera.
It also influences economic policy by governments, by central banks, and of course, information has a profound impact on the functioning of financial markets.
Of course, not all information that we would like to know is actually known, and this is reminiscent of the framework that was proposed by Secretary of Defense Donald Rumsfeld back in 2002, when he spoke about the known knowns and the known unknowns, et cetera et cetera, as shown in the exhibit that is appearing on the screen now. With this framework in mind, how may artificial intelligence change things? A useful starting point is the observation that information gathering and information processing may be expensive and certainly is time-consuming.
Artificial intelligence can change this very significantly to the extent that the necessary and expensive IT investments have been made or that the technology has been bought.
Let's now start with the known knowns. Here, AI can make a big change because it will enable the user of the technology to analyze more data more quickly and more cheaply. This means it has a profound impact on the way that companies, firms, governments, et cetera, individuals, will use and access information. One can think about the production of PowerPoints, producing summaries, analyzing texts, analyzing data for marketing purposes, for credit risk analysis, for financial market investments, et cetera, et cetera. The unknown knowns concern information of which we know that they exist but that until now has not really been used because the cost was prohibitively high.
The cost involved the use of lots of analysts, manpower, and that meant actually that such an investment would have delivered a very low return, could even have delivered a negative return on investment. When you introduce AI and its capacity to go through lots of data quickly and cheaply, the economics change completely and it means that the return on investment of tapping into data that hitherto have not been used becomes attractive.
The next box concerns the known unknowns. This is actually an interesting case. Why? Because we know that there is a relationship between, say, economic variables, between A and B, but we do not know the exact nature of the relationship, what is driving it, how one is influencing the other, and so on.
So here again, artificial intelligence can change things very significantly because of the data processing capabilities, because of the possibility to generate actually synthetic data by combining proprietary data of a company, say, published data, say macroeconomic time series, but also data that you get from the internet or even analysis of what is circulating in terms of comments on social media.
So what follows is actually an understanding, a possible understanding of how A influences B or vice versa. This is kind of a new world in which we enter. The last box concerns the unknown unknowns.
This is actually a very fascinating box because it concerns topics that going forward may influence the activity at some point, but of which we are unaware and we are even unaware that we ignore these topics, that we do not know these topics. So how do you cope with this?
This is actually a very interesting question from a company perspective or even from a public sector perspective. Well, here again, artificial intelligence can offer us some sort of solution by, again, going through the data and actually discovering relationships of which we were even unaware that they existed.
To conclude, what does all this mean in terms of potential economic consequences? Well, focusing at the level of a company, the government, the central bank or an asset manager, there is the element of IT costs, the cost of developing or of buying a technology.
There is the element of buying the access to the data. This is an important element to be kept in mind. And thirdly, there's also the element of the human skills, which are necessary to use a technology to think about how it can be used to tap into the available information.
When you do all these things, when you do them right, you can expect to be better in doing your job, to have more ideas, to have a better marketing strategy, better pricing, but also to realize productivity gains at the level of the individual company.
When you look in the aggregate, there are, of course, many, many consequences, but one specific consequence to highlight is, of course, the disruption that will follow in the labor market and which clearly will create a challenge for governments to kind of accompany that disruption and to take the right policy initiatives to facilitate moving from one skill level to another skill level to be able to cope with these new technologies.
Another consequence is that you would expect that it would increase competition, because competition rooted in the fact that you have more data, you access the data more quickly, you come to better conclusions.
But it also raises the question whether it could also lead to more concentration later on, more concentration because of the fact that within a given sector, not every single company will be sufficiently agile, will have the necessary people with the right skill set, will have the financial means or will have the vision to actually reap the benefits that AI may bring.
So, lots of change is upon us in different aspects, in different directions, and we will cover that, of course, going forward in the work of our economic research team. I thank you for having watched EcoTV and please join us again soon. Thank you.
