The US authorities’ decision to ban non-US citizens from using two pieces of software developed by Anthropic has sent shockwaves through the world of artificial intelligence.
Firstly, because it is unlikely that it can be enforced; secondly, because it reflects a profound misunderstanding of the science and its dissemination; and finally, because it is completely ill-suited to addressing potentially legitimate concerns about the immense new capabilities for computation and data aggregation being used for criminal purposes.
The United States strongly criticised Europe for attempting to regulate the tremendous scientific advances in this field based on a few simple principles, starting with the idea that we cannot accept on the internet what we forbid in real life.
They now find themselves with their backs against the wall, forced to ban the use of software — admittedly with unprecedented capabilities —‘for reasons of national security’. What a lesson!
The software’s own creators and owners had duly warned against this and refused to allow the military to use it for that purpose. They were punished by the very same administration, which proves that it intends to reserve these scientific advances for itself, to turn them into a weapon.
Yet this will not be the case. No coercion will halt the frenzied race set in motion by AI, save perhaps for ethical and moral reasons.
Is it not time for international consultation at the highest level, involving all those who are or will be affected by this issue?
For its part, the European Union must strive to learn from this. Too much regulation could hinder the emergence of European leaders in this field, but it is not always the regulation one might imagine.
The regulation concerning the principles protecting the individual from commodification or the violation of their integrity cannot be contested. Increasingly shared, including elsewhere in the world, it must be preserved and protected.
On the other hand, the myriad administrative obligations that prevent start-ups from taking risks and finding the financial means to get off the ground must be systematically eliminated in Member States as well as at European level.
Europe’s innovation gap is not a scientific or technical one; it is that of a society which seeks to minimise risks at all costs and is therefore unable to provide the financial resources for the development of new technologies—resources that the US stock market makes available to those who venture to take risks, even if it means losing.
Elon Musk has just capitalised on this by raising $75 billion on the New York Stock Exchange by promising Mars and the Moon!
Without going that far, Europeans must obviously, and more than ever before, step up their support for innovation. They still have work to do: artificial intelligence at the cost of risk; intelligence for the common good, period!
