Alignment. Alignment. Alignment. It might be the word of the year in AI. Everyone talks about aligning AI with our values and goals. But aligned with whom, and toward what?
On one side of the world, a state increasingly treats AI as national infrastructure, integrating it into industry, science, education, healthcare and public services. On the other, another state openly frames AI as a race for global dominance, economic power and military advantage.
Both can build systems that are perfectly “aligned.” But aligned to very different things.
We talk about alignment as though humanity has already agreed on the destination and the only problem left is making the machine follow the road. We haven't. An AI can be aligned with a country, a military, a corporation, its users, or supposedly “humanity” — but even then, who decides what humanity wants?
So perhaps the real question isn't simply whether AI is aligned. Whose goals are we putting into the machine?
[link] [comments]