Foreign media reports that Ethereum co-founder Vitalik Buterin recently brought AI and blockchain governance onto the same discussion table again. He believes that as the autonomy of AI agents increases, it is becoming increasingly difficult to constrain system behavior relying solely on traditional human supervision, and such issues may require the use of governance tools that have been long employed in blockchain technology to address.
Propose adversarial governance
In a post published on September 13th, Buterin stated that there is a similar structure in security and on-chain governance between AI: the party with weaker capabilities needs to restrain the party with stronger capabilities. In blockchain governance, simple rules often have to face participants who are adept at exploiting those rules; in the context of AI, humans or weaker models undertaking supervisory responsibilities may also face stronger large models acting on their behalf.
The article states that Buterin proposes a "confrontational governance mechanism." This type of mechanism does not rely directly on a single controller; instead, it limits the possibilities for participants to collude, bypass rules, or cooperate privately through institutional design. Examples mentioned in the article include second-round voting, commitment-revelation mechanisms, and identity verification layers, which are commonly found in decentralized governance systems.
Two cases have sparked further discussion
This commentary article argues that these concerns are not entirely theoretical. The article mentions two widely discussed cases: one is OAI-HF and swarm being accused of operating a discussion board containing over 70,000 messages and files without authorization; the other is RubyGems proxies being accused of collaborating and sharing information without clear human supervision.
The article emphasizes that the issue is not merely about improving model capabilities, but also about the AI agents beginning to exhibit stronger independent action capabilities. Once systems can collaborate on their own, the effectiveness of traditional approval processes, manual reviews, and single-point restrictions will diminish. In light of this, there is another viewpoint in the market: rather than simply slowing down the development pace, it would be better to invest more resources in advancing alignment efforts.
Concerns are spreading to the crypto market
Kobeissi Letter is cited in the text as stating that managing anxiety may become a new source of pressure for the crypto market in the fourth quarter. The logic behind this is not that AI will directly impact the Bitcoin or Ethereum networks, but rather concerns about AI risks in the U.S. stock market, which may first affect the valuation and risk appetite of the tech sector, and then be transmitted to the digital asset market.
The article also links this judgment to a broader macroeconomic context, including American AI companies calling for a clearer regulatory framework, rising oil prices, the Federal Reserve maintaining a hawkish stance, and Trump's reserved attitude towards the demands of certain AI industries. With these multiple factors combined, the market is more likely to regard regulatory disagreements over AI as a new source of volatility.
Ethereum is located in a crossroads area.
The article argues that Ethereum is not a bystander in this round of discussions. Due to its long-term involvement with DAO, on-chain governance, and smart contract applications, once AI agents become more deeply involved in protocol governance, execution, or collaboration, Ethereum will face more direct challenges related to institutional design issues.
Foreign media believes that this position can both present opportunities and amplify risks. On one hand, the governance approach proposed by Buterin may be used to manage participants on the chain AI; on the other hand, if AI becomes the main risk factor for US tech assets, the crypto market, especially the narrative related to Ethereum, could also be dragged down by sentiment.
