Against AI & crypto

A lot of people hate crypto so let's come up with arguments against the idea that AI agents will use crypto to facilitate payment for transactions

Introduction

My most recent post argued that autonomous AI agents transacting with each other will use crypto as a medium of exchange. And I noted that a lot of people are axiomatically opposed to crypto, and so they react negatively to this suggestion. Well, a bunch of people unsubscribed to this Substack after I published that post, so I guess it’s really true that people have strong feelings about crypto.

Let’s take the opposite argument. What if we come up with some reasons why AI agents autonomously transacting with each other won’t use crypto to facilitate payment? I’ve come up with eight plausible counterarguments to my earlier post. None of these counterarguments really convince me that my initial argument is wrong.

Regulatory Barriers

While crypto is currently unregulated enough to allow AI agents to use it for transactions, that could change. Governments are paying more attention to crypto, and they could impose regulations that limit the ability of non-human entities to hold or transact in cryptocurrencies. Such regulations would obviously limit the ability of AI agents to pay for transactions with crypto.

Counterpoint: Regulation lags behind innovation. By the time governments regulate this area, AI-driven crypto usage will be too widespread to fully control.

Security Concerns

AI systems are vulnerable to hacks, data breaches, and manipulation. Without robust security measures, AI agents transacting with crypto might enable large-scale theft. Additionally, as AI agents become more complex, it will be harder to audit or regulate their behavior, which makes accountability difficult.

Counterpoint: These risks exist for human actors, as well, and technological improvements in AI could mitigate these concerns over time.

AI Autonomoy Might Be Overestimated

A major assumption is that AI agents will become autonomous enough to transact without human oversight. However, we are still far from creating truly autonomous AI agents which can execute financial transactions independently. Much of the autonomy that we see still requires significant human input, programming, and oversight. AI agents that execute transactions without human oversight may not materialize for a while.

Counterpoint: Technological progress can surprise us, and sufficiently autonomous AI agents might arrive faster than expected. Early use cases provide an indication of the direction in which this technology is headed. The question is how quickly the future arrives.

Crypto Volatility Makes it a Poor Medium of Exchange

Cryptocurrencies are notoriously volatile, which makes them a poor medium of exchange. This volatility creates transactional inefficienies, espeically for larger transactions. Stablecoins like USDC solve this problem, but they rely on trust in the crypto ecosystem, which, as I indicated earlier, is not universally loved.

Counterpoint: Stablecoins and other financial instruments can reduce volatility, and their adoption is growing. Crypto volatility may become less of an issue as markets mature.

Human-Centric Economy

The global economy is structured around human decision-making and agency. The idea of AI agents transacting autonomously could disrupt this to an uncomfortable degree, but there might also be structural limits to how much economic activity can realistically be automated by non-human entities. Many economic processes still rely on human creativity, intuition, and decision-making. Autonomous AI agents might excel at certain kinds of transactions, but large-scale adoption may be far off in industries that are deeply human-centric, like law, healthcare, or education.

Counterpoint: Automation has already displaced jobs in these sectors, and AI will likely erode more roles over time. A hybrid model of human-AI interaction could emerge in many areas, including economic transactions.

Public Distrust and Backlash

The public’s reaction to AI and crypto is crucial. If the people who unsubscribed to this Substack after I published my most recent post are any indication, a lot of people hate the idea of autonomous AI agents using crypto to facilitate payment for transactions. If the public thinks that AI agents are too autonomus, political movements aimed at stopping these developments will arise. Public policy is often shaped by public sentiment, and a populist response to these concerns would be an obstacle to full adoption of autonomous AI agents transacting with crypto payments.

Counterpoint: While backlash is possible, or even inevitable, most innovation occurs despite public opinion, with policymakers following rather than leading. By the time people realize the full impact, AI will already be entrenched.

AI Agents May not Need Crypto

An alternate possibility is that AI agents won’t require crypto to transact. Instead, traditional finance systems could evolve to accommodate AI. Banks could develop AI-centric products that allow agents to participate in the economy in ways that comply with human regulations. Such innovations could render crypto less neccessary in the long run, particularly if traditional financial institutions figure out how to safely integrate AI transactions.

Counterpoint: Traditional financial systems are notoriously slow to evolve, whereas AI is rapidly improving.

Technical and Economic Feasibility

Maybe it’s overly optimistic to assume that AI agents will be sufficiently autonomous to execute financial transactions without human intermediaries. Further, AI systems require significant computational power, maintenance, and resources. The value of autonomously transacting AI agents might not justify the costs involved. AI agents might still need to rely on humans for certain inputs that would prevent full autonomy.

Counterpoint: As computing costs decline and AI infrastructure improves, the feasibility of autonomous economic activity will be less constrained by these factors.

Conclusion

While there are valid and significant challenges to the idea of autonomous AI agents using crypto for transactions, none of these counterarguments convincingly negate the trajectory we’re on. Regulatory hurdles, security risks, and public skepticism are real, but innovation tends to push ahead despite such barriers. Crypto’s volatility and the limitations of current AI autonomy may slow adoption, but technological advances in stablecoins, decentralized security, and AI sophistication suggest that these issues will be solved sooner than many expect.

It’s worth considering the alternative: Can traditional financial systems evolve quickly enough to accommodate autonomous AI transactions without the need for crypto? History suggests otherwise. Legacy systems move slowly, weighed down by bureaucracy and inefficiency, whereas decentralized crypto networks offer flexibility and speed that align with the inherent nature of AI autonomy.

Ultimately, while I’ve explored counterargumetns that question the integration of AI and crypto, none of these challenges feel insurmoutnable. The march toward AI-driven, autonomus transactions is likely inevitable, and crypto—with its decentralized, trustless architecture—seems well-positioned to serve as the backbone of this future. It may not happen overnight, but the alignment of these technologies suggests that autonomous agents will participate in the global economy, unhindered by the limitations of the traditional financial system.