If I expected that an AI above a certain line will be nationalized, I would try to get just below than line and stay there for as long as possible, to maximize my profits. Alternatively, I may choose to cross the line at the moment my competitors get close to it, if I want to be remembered by the history as the one who reached it first.
If I expected that an AI above a certain line will be nationalized, but I believed that the government will give a lot of money in return (for example, to convince other entrepreneurs that the country is not turning into a communist dystopia), I might decide to try to get that money as soon as possible (less work, guaranteed profit, a place in history), so I would actually exaggerate how dangerous my AI is, to make the government take it away sooner.
But if my goal is to avoid nationalization...
One option is to make my work distributed across countries, so whenever one of them starts talking about nationalizing it, I will make it clear that they can only take a small part of it, and I will simply continue developing it in the remaining countries. I would move the key members of my company to countries with the lowest probability of nationalization. Actually, they could live in a country where I have no servers, and they would connect to them remotely. So that when the government takes the servers, it cannot compel the people to explain how the entire thing works, or prevent them from remotely destroying the part that got nationalized. Also, every part would have a backup in another country.
Another option is to make the AI pretend that it is less smart than it is, to officially stay below the line. It would mean I cannot directly sell its abilities to customers, but maybe I could use it myself, e.g. to let it manage my finances, or I could use it in a plausibly deniable way, e.g. you can hire my smaller company that hires 100 experts who also use my AI, and everything they do is officially attributed to the genius of the human experts.
If I expected that an AI above a certain line will be nationalized, I would try to get just below than line and stay there for as long as possible, to maximize my profits. Alternatively, I may choose to cross the line at the moment my competitors get close to it, if I want to be remembered by the history as the one who reached it first.
If I expected that an AI above a certain line will be nationalized, but I believed that the government will give a lot of money in return (for example, to convince other entrepreneurs that the country is not turning into a communist dystopia), I might decide to try to get that money as soon as possible (less work, guaranteed profit, a place in history), so I would actually exaggerate how dangerous my AI is, to make the government take it away sooner.
But if my goal is to avoid nationalization...
One option is to make my work distributed across countries, so whenever one of them starts talking about nationalizing it, I will make it clear that they can only take a small part of it, and I will simply continue developing it in the remaining countries. I would move the key members of my company to countries with the lowest probability of nationalization. Actually, they could live in a country where I have no servers, and they would connect to them remotely. So that when the government takes the servers, it cannot compel the people to explain how the entire thing works, or prevent them from remotely destroying the part that got nationalized. Also, every part would have a backup in another country.
Another option is to make the AI pretend that it is less smart than it is, to officially stay below the line. It would mean I cannot directly sell its abilities to customers, but maybe I could use it myself, e.g. to let it manage my finances, or I could use it in a plausibly deniable way, e.g. you can hire my smaller company that hires 100 experts who also use my AI, and everything they do is officially attributed to the genius of the human experts.
We had a similar thought:
But yeah, my initial comment was about how to take advantage of nationalization if it does happen in the way Leopold described/implied.