morpho logo bg
Co-founder Paul Frambot says the unauthorized tweet was generated by a third-party tool, access to which has been revoked
A DeFi lending protocol’s official account briefly told the world that its own business model runs on backroom deals. Morpho says a rogue AI marketing tool is to blame.
On September 25, a post published from Morpho’s official X account described the protocol’s curator businesses as dependent on private, off-chain distribution deals rather than the transparent on-chain vault mechanisms the project has spent years building. The post was deleted shortly after, but not before it caught the attention of the DeFi community.
What actually happened #
Morpho co-founder and CEO Paul Frambot addressed the situation on September 26, confirming that the tweet was not written or approved by anyone on the team. Instead, it was generated and published autonomously by a third-party AI marketing tool that had access to the protocol’s official X account.
Frambot said access to the tool was revoked immediately and that the team launched an investigation into how the post was triggered. The specific AI marketing tool involved has not been publicly identified, and no details about what prompted the post have been disclosed.
How Morpho’s curator model actually works #
To understand why the deleted post stung, it helps to know what curators actually do within Morpho’s ecosystem. Curators are independent entities that manage on-chain lending portfolios built on top of the protocol. They design vaults with specific risk parameters and earn fees based on the performance and structure of each vault they manage.
The system is designed to be transparent by default. Morpho’s Vault V2, which formally launched in September 2025, introduced a role-based architecture that gives curators granular control over risk settings. Critically, these actions are subject to timelocks, meaning changes to vault parameters aren’t instantaneous. There’s a built-in delay that allows depositors to review and react before modifications take effect.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
Morpho had been actively expanding its curator tooling in the months leading up to the incident, with a stated focus on enhancing transparency and improving risk management capabilities.
The broader AI-in-DeFi problem #
The use of AI tools to manage social media accounts is not unique to Morpho. Across the crypto industry, teams have increasingly adopted automated content generation and scheduling tools to maintain a consistent posting cadence. The risk, as Morpho just demonstrated, is that automation without sufficient oversight can produce messages that are factually wrong or strategically damaging.
What makes this case notable is the specificity of the error. This wasn’t a generic marketing tweet with a typo. The AI tool generated content that made a substantive, incorrect claim about how Morpho’s core business model functions. That suggests either the tool was drawing on outdated or inaccurate training data about the protocol, or its content generation logic lacked the guardrails needed to prevent publishing claims that contradict a project’s fundamental architecture.
What this means for DeFi protocols #
For Morpho specifically, the damage appears contained. Frambot responded quickly, the post was deleted, and access to the offending tool was cut off. The investigation is ongoing, but the immediate crisis management followed a reasonable playbook. Protocols that grant AI tools publishing access to official accounts are effectively outsourcing a piece of their credibility to software they don’t fully control. For investors and depositors evaluating DeFi protocols, the incident adds another dimension to due diligence. Understanding how a protocol manages its operational tooling, including who or what has access to its official communications channels, now matters alongside traditional evaluation of smart contract architecture and tokenomics.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our