Pure extractive progress?

There is this parallelization that data is the contemporary oil (which in its turn was the new gold) due to its importance to modern economies. And as with all analogies, the mind tends to stretch them to their full extent. So do I.

In the center of gold, oil, and now data extraction and profitability machines lies the scarcity of these resources—they are not infinite, and this means that the one who controls access to them has power due to the inability of others to act independently. Consequently, it heavily incentivizes those having the “extraction rights” to block any other from entering their field—both by controlling access to the now-known resources and also by trying to capture (and block access to) any newly or independently held resources. And this analogy can bring us to the era of AI as a (somewhat) new and glorious extraction territory.

If you prefer, you could picture big AI models and their respective owning legal entities as modern looters—they just burn as much and as quickly as they can afford from the new oil in order to maximize extraction profit. What remains is then similar to an exploited and abandoned gold mine or oil station. The rate of exploitation (and depletion, consequently) in the case of contemporary AI is such that the above seems a valid remark. What a misuse of precious resources that could be? What a lost opportunity to collectively prosper and progress.

But then data are intangible, can exist in many copies, and can be altered and reused—their digital nature may be the one that will save them as they cannot be held—at least not in the way gold and oil could. And in that way no one can truly deplete them but merely use some of their stored value before they cycle back in a new form with more value and be available for another round of extraction—hopefully more beneficial, just and equally distributed.