Guiding rather than containing
There is much talk about the risks associated with AI and the innovations that could emerge from it. As a result, the watchword of the European wise men, who are well versed in totalitarianism of all kinds and therefore sensitive to surveillance issues, is caution and anticipation.
However, several challenges make this excess of caution outdated and, a priori, counterproductive. To begin with, latency and unpredictability.
The first reason why it is futile, roughly speaking, to regulate AI as the Europeans are trying to do, is latency. Already mentioned by Mac Bloch in L'Etrange Défaite, or by the author in writings on Quantum Management (Arreola, Cartau), the distance constituted by the bureaucratic hierarchy, and therefore time, is far greater than the time it takes for innovation by AI to progress. This must be remembered.
Marc Bloch, already, recounted how the military hierarchy of the First World War had had time to adapt to technological and organizational innovations. Four years had left room for a general staff far removed from the theater of operations to adapt.
This was not the case during the blitzkrieg. As the name suggests, there is no time for adaptation.
Furthermore, as mentioned in Quantum Management, observation and action are so closely linked that they are sometimes inseparable. This is all the more true given the rapid pace of innovation.
In this sense, European regulations are akin to a Maginot line that has been largely overtaken by mobility, digital in this case.
The other argument, to dispel any faint-heartedness, lies in the unpredictability of innovation.
Qwant now offers a very nice AI tool for synthesizing search results. In my prompt, “Managing the Societal Risks of Innovation”, the recommendation is made in particular to “assess its potential impacts on society...before launching an innovative project” because “it is essential” or, “involve stakeholders” or even, “make sure that your innovations comply with laws and regulations”.
However, when innovating, even if you set yourself a goal and anticipate the broad outlines of its compliance, the functional and viable result will very often be different from what was planned.
For example, I discovered this morning that Grok could help me explain to my 12-year-old son how a quantum computer works. What a godsend! Imagine all the teaching aided by AI! At my level, I am just starting to adapt to innovations. But had we anticipated this result, or the impact it will have on the teaching profession? What would the reaction of the EN be if I involved it as a “stakeholder”?
On a larger scale, more profound innovations are in the works. And in concrete terms, only an AI will be able to adapt to another AI and above all to what we cannot foresee, even a few weeks in advance. We must play on equal terms and only a level of technological competence within sight can the one that precedes it.
When Gutenberg invented the printing press as we know it, the Catholic Church was unable to suppress or regulate it. It had to adapt. It wasn't necessarily for the best, but it allowed it to pursue its mission.
In its early days, Google had no idea what its search engine would become. So much so that it adopted the motto “Don't be evil”...
The risks of surveillance, you might ask? They are not so much technical as institutional. In a fully functional legal framework, surveillance and eavesdropping are governed by laws that are respected.
It is hubris, even misplaced contrition for Europe to think that it will save the world through its prudish virtue.
I believe it was Dale Carnegie who told the story of how he decided to keep a manager who had lost $1 million, which was no small amount at the time. When asked why he hadn't fired the manager, Carnegie replied: “It was a very expensive lesson. Surely you don't think I'm going to fire him once the lesson is over!”
Europe has learned a lot and we must not throw the baby out with the bathwater. Its moderation will only be effective in guiding, not containing. Innovation can only be guided, it cannot be stopped. Europe must rely on businesses and its tens of millions of enlightened minds to guide its destiny, not a few large piles of paperwork that are outdated even before the printing is finished.
Otherwise, Europe is condemning itself to a Strange Defeat.
Philippe Cartau
Comentarios