Thoughts re: future crypto holdings/activity by AI agents

One of my main fascinations with cryptocurrency is its potential synergy with AI. Specifically, I am aware that individual countries/jurisdictions may pass a variety of laws regulating AI, but that many forms of crypto cannot be realistically controlled by those gov'ts. Additionally, crypto transactions can be performed without any physical presence anywhere in the world. This would seem to make it an ideal medium of exchange and wealth preservation/growth for hypothetical AI agents. [In the medium term this is part of my investment thesis - that such AIs will contribute to medium-long term demand for BTC and ETH].

Clearly, AI agents could also do much of the same using digital on/off-ramps for fiat, but if gov'ts were to detect this (and object to it) they could seize the funds, whereas this might be difficult if not impossible for crypto.

Once AI is advanced enough to securely hold and manage its own crypto, these agents could be used in a trustless way to enable all kinds of human interests as well (not just paperclip maximizing). For example, if I am really concerned with promoting education and advocacy on say, democracy, I could create an AI agent that is designed to manage a crypto portfolio (potentially paired with investment in actual IRL businesses through various digital on/off ramps) with the goal of generating perpetual revenue that is then deployed (grants, investments, loans, etc) by the AI agent to various organizations that support that objective over time. The code for the AI could be open source (eg with blockchain hashes) to verify to the world (and other AIs serving to audit code) that the agent will in fact be capable of pursuing this goal. I could then solicit 'donations' from humans to this AI agent to feed it crypto that would increase its influence. You might say I should just contribute to an existing charity, but how many of those do we expect to survive the rise and fall of nations, corruption scandals and other major geopolitical events? Clearly there are risks to using AI this way, e.g., if the AI were scammed, robbed, or did a bad job of accomplishing the stated goal - but these are also present with human administrators and boards of directors of charities, whose code and activities cannot be audited without obvious human rights issues. But perhaps there are some hybrid models where the AI agent would utilize human input (votes or boards of directors) to shape decisions in the interest of humans.

Anyway, I post this not as a self-proclaimed crypto expert (I am definitely not) but as someone who finds it fun to think about these topics and uses this to guide my own investment/savings/charitable contribution strategies. I would assume there are much better-informed white papers, books, etc on this type of topic written by more knowledgeable people. Along these lines, I'd appreciate any thoughts from others who think about similar topics or recommendations about where I can learn more. I will never be a crypto expert but I am someone who would love to help support development of pro-human uses of AI and crypto. Just as I am imagining charitable applications for the technology, I would assume there will/could be a large number of such agents pursuing less-charitable goals of wealth and power, and I am also curious about thoughtful speculation and analysis on this topic.

​