Description du poste
I've been increasingly concerned with how international law has been sidelined as a potential solution to global, international problems, AI safety being one of them.
In the past few years, various governments in western countries have shifted the Overton window on international law, making it seem irrelevant or useless. Researchers have now accepted this shift without challenging it, in a dangerous and negligent manner.
A decade ago, it would have been unthinkable to call the UN "useless" or "irrelevant." Now that's a common sentiment, even among senior researchers in AI safety.
Researchers at Oxford University for example, present the prototypical case. Many were researching law and international law solutions for decades, but in the last decade, many have pivoted to focusing exclusively on technical AI governance, almost as if their legal background doesn't exist. This is merely an observation of published papers from the leading institutes. I am not intending to be rude to individuals, just pointing out the very clear trend lines.
Why international law mattersThe current miopic view of the UN and international law is false, ahistorical and dangerous because it sidelines one of the most effective tools in human history.
International law is a major cause of peaceTo refute the "UN bad" sentiment, it's important to understand the context to truly get what the UN has achieved. Prior to WW2, Western European countries were at war on and off for hundreds of years, with it rarely being more than a few years between wars.
In WW1, Woodrow Wilson came up with the idea of the League of Nations, a place for countries to meet to resolve disputes. It ultimately failed in part because it did not include Germany. (The lesson being that everyone needed a seat at the table).
After WW2, the UN was created and a flurry of international agreements, laying the foundation for our modern world, on everything from borders and trade to human rights, airline safety etc.
Since the end of