What struck me is that the hardest questions in your examples are not actually about AI. AI can help aggregate preferences, reveal common ground, and improve deliberation. But when the question becomes who gets to vote, who bears the costs, and who gets to decide, we are no longer dealing with an information problem—we are dealing with a power problem.
In a sense, your wartime examples reveal the limits of digital democracy not because AI is insufficient, but because democratic procedures eventually run into underlying power structures.
This connects strongly with a theme I’ve been exploring in my Rethinking the State series: states are often treated as moral actors, but they may be better understood as power-organizing systems. AI might help us hear society more clearly, but it does not tell us how power should be allocated when different groups bear radically different costs.
Looking forward to reading more of your work on AI, power, and governance.
Thank you so much for your thoughtful comment! I completely agree. AI tends to amplify and accelerate the tendencies that are already present in a system. If digital democracy already suffers from issues of fairness or justice, introducing AI without extreme care will only compound those flaws.
Ultimately, implementing AI requires us to look closely at power dynamics and broader societal impacts, not just the technology itself.
I've just subscribed to Deep Bitcheese Brew. I'm really looking forward to your pieces.
I suspect many of us are arriving at a similar realization from different directions: the most important AI questions may ultimately be governance questions.
Really enjoyed this piece, Taras.
What struck me is that the hardest questions in your examples are not actually about AI. AI can help aggregate preferences, reveal common ground, and improve deliberation. But when the question becomes who gets to vote, who bears the costs, and who gets to decide, we are no longer dealing with an information problem—we are dealing with a power problem.
In a sense, your wartime examples reveal the limits of digital democracy not because AI is insufficient, but because democratic procedures eventually run into underlying power structures.
This connects strongly with a theme I’ve been exploring in my Rethinking the State series: states are often treated as moral actors, but they may be better understood as power-organizing systems. AI might help us hear society more clearly, but it does not tell us how power should be allocated when different groups bear radically different costs.
Looking forward to reading more of your work on AI, power, and governance.
Thank you so much for your thoughtful comment! I completely agree. AI tends to amplify and accelerate the tendencies that are already present in a system. If digital democracy already suffers from issues of fairness or justice, introducing AI without extreme care will only compound those flaws.
Ultimately, implementing AI requires us to look closely at power dynamics and broader societal impacts, not just the technology itself.
I've just subscribed to Deep Bitcheese Brew. I'm really looking forward to your pieces.
Thank you, Taras.
I suspect many of us are arriving at a similar realization from different directions: the most important AI questions may ultimately be governance questions.
Looking forward to continuing the conversation.