There's a deep failed idea here. It's the idea that technology is an alternative to politics, period. I think that whole idea has failed.
You can't avoid building technologies that could be used by a police state, because literally everything can be used for warfare or aggression. The first person who sharpened a stick or a rock created a weapon. Everything is a weapon. Everything is dual-use.
If you want to prevent a police state or any other undesirable political outcome, you have to confront the politics directly. "Change the tools and you change the rules," an old saying from 1990s cyber-culture, is an idea that I think has failed. Instead what happens is that the rules determine what's done with the tools. Those 90s cyber-hippie and cyber-libertarian types vastly underestimated the power of the political to the point of making a category error.
Cryptocurrency is a great case in point. It started as an attempt to create a technology that would act as an escape hatch from corrupt casino capitalism and state-backed financialization. Instead it became an insane absurdist parody of those things. The same types of corrupt individuals who profited off "too big to fail" and the rest of it simply took cryptocurrency and used it to implement the same type of system but on a block chain. The rules (and rulers) changed the tools, not the other way around.
I mean you kind of can. The whole premise of AI is basically that we'll make white collar administrative compliance work dirt cheap in the long run and we'll all be richer for it and we'll sidestep the whole "fight with all those industries" step because the sheer economics of it will make it happen and as those industries become cheaper/poorer there will be even less reason to not kick them to the curb. Kind of like how once certain state shed a lot of their manufacturing or agriculture, adopting policy that's bad for the rest became easier. Or how now that solar is cheap we mostly DGAF about coal.
Not everything is dual use. Some information technology grossly favors centralization or decentralization of power based on the type of information problem it solves.
Cryptocurrency is clearly a counterexample to your main point. The fact that some people use cryptocurrency tech to implement centrally managed private currencies is largely irrelevant. To the extent that cryptocurrencies are used to implement private currencies, it largely just hampers them technolgicially if it has any effect at all. These projects just like to use the term cryptocurrency to imply some level of legitimate decentralization or security that doesn't really exist.
It's like saying that because you can use a screwdriver as a weapon, it's inherently a destructive technology (or some mixture of destructive and constructive). On closer inspection it's clear that the screwdriver doesn't sit on the pareto front for weapons, but it does sit on the pareto front for manufacturing tools. So a screwdriver is clearly a technology whose existence favors construction over destruction. Similarly I would argue that technology like bittorrent or bitcoin favor decentralization over centralization.
As another example, surveilance technology favors centralization. None of what we are seeing now with flock would be possible without cheap mass produced digital cameras. The balance of technology favors cheap cameras (and counter-surveilance is often impossible or expensive) and it favors centralization because you need a lot of data to create inferences based on this spurrious data.
If we could have predicted these effects, perhaps we could have delayed the development of cheap cameras, but it just goes to show how complicated this game is. We could just as easily have said that more cameras would lead to unprecedented public transparency (e.g. bodycams on cops, dash cams). Unintended consequences.