Eliza Labs founder Shaw Walters has sent a clear message that cuts through the excitement around artificial intelligence. According to him, AI agents are not yet capable of safely managing people’s money.
In an interview with Decrypt, Walters explained that while AI can process information faster than any human trader, it still lacks judgment, intuition, and emotional understanding.
He believes the technology is more useful today as an assistant that helps structure information and improve decisions rather than as a replacement for human expertise.
Eliza Labs Founder’s Thoughts on AI and the Future of Financial Agents
When Eliza Labs launched ElizaOS earlier this year on Solana, the team introduced an open system where developers could create intelligent agents and simulations.
Source: SingularityNET
Walters clarified that this was never meant to be a tool for fully automated trading. Instead, he said the goal was to create technology that makes markets more transparent and trustworthy.
One of Eliza’s most interesting innovations is what Walters calls a marketplace of trust. The platform uses AI agents to simulate token purchases and test the credibility of different users. “We can identify who’s actually good at calling and who’s trying to scam,” Walters explained.
The system “paper-buys” tokens to evaluate whether someone’s predictions are accurate, building a reputation score for every participant. This approach helps distinguish honest contributors from manipulators, creating a fairer environment for community-driven analysis.
Eliza is also experimenting with an agent-run over-the-counter desk, where users can negotiate token purchases directly with AI agents. However, each transaction operates within clear boundaries set by smart contracts.
These limits ensure that every interaction remains transparent and prevents the system from acting beyond its safe range. Walters described this as an early model for responsible autonomy, one that combines the speed of AI with the discipline of human-defined rules.
He also spoke about Eliza’s upcoming developments, including a token migration and cloud rollout designed to expand its governance features and allow agents to function across multiple blockchains.
The new architecture aims to make it possible for users to deploy their agents with a single command. Even so, Walters remains firm that the goal is not to hand over control of finances to machines.
“Fully autonomous trading agents are not investment-ready,” he said. “They can be fast, but they do not understand the context or consequences of their actions.”
Beyond the technical discussion, Walters is also facing a legal battle. In August, Eliza Labs filed a lawsuit against Elon Musk’s X platform.
The company claims X obtained confidential details about Eliza’s technology under false pretences, then used the information to launch a competing product while banning Eliza from using the platform.
Walters accused X of damaging Eliza’s reputation and blocking access to customers and investors. The case highlights the challenges faced by independent AI innovators competing with powerful corporations.
Despite these setbacks, Walters continues to focus on building tools that support collaboration and insight rather than control. He described the role of AI in finance as one that should guide, not dominate.
“It’s more about recommendation and insight and helping with community coordination than blindly following the AI,” he said. This reflects a belief that technology should enhance human intelligence, not override it.
The Risks of Letting AI Manage Money in Crypto
Walters’ warning carries real weight in the context of cryptocurrency markets. Unlike traditional finance, crypto moves without pause and often reacts to sudden waves of sentiment.
Prices can rise or fall within minutes, and even a small miscalculation by an autonomous system could cause massive losses. AI may process data faster than any human, but its lack of context can turn that speed into a risk.
One of the main problems is that AI relies entirely on patterns found in historical data. Yet crypto markets are deeply shaped by unpredictable human behaviour, changing regulations, and cultural movements that no algorithm can fully interpret.
A tweet from a well-known figure or a rumour spreading online can completely change the direction of a token’s value. While AI can detect that movement, it cannot understand why it happened or whether it will last.
Accountability adds another layer of concern. If an autonomous agent loses funds, who is responsible?
The investor who used it, the developer who created it, or the blockchain protocol hosting it? With no clear answer, the financial and legal risks remain significant. That is why Walters believes human oversight must always be part of any financial system involving AI.
Security risks also play a major role. An AI agent is only as safe as the code behind it. If a vulnerability exists in the logic or in the connected smart contracts, it can be exploited by attackers.
In decentralised finance, where vast sums of money move through automated systems, such weaknesses can lead to devastating consequences. AI systems that depend on external data feeds are especially vulnerable to manipulation, a problem that continues to challenge developers.
Beyond the technical issues, there is the problem of emotional intelligence. Although AI never panics or becomes greedy, it also cannot sense danger when markets turn chaotic.
Human traders often make decisions based on instinct or fear, which can prevent deeper losses. An AI agent, on the other hand, acts purely on logic, which can be dangerous when the data it receives does not reflect the full reality.
Walters’ approach treats AI as a guide that simplifies complexity rather than a machine that acts alone. Eliza Labs is building technology that helps humans see patterns, organise data, and coordinate actions more effectively.
The company’s approach shows that the future of finance lies in collaboration between people and machines, not in one replacing the other. AI can analyse, suggest, and even execute, but it still needs human judgment to steer it through uncertainty.
Conclusion
Shaw Walters’ perspective offers a grounded view amid the growing noise surrounding AI trading. His message is not one of fear but of balance. Technology can be a powerful ally in finance, but it cannot replace experience, caution, and emotional understanding.
Eliza Labs’ work on Solana shows that progress in AI does not have to mean surrendering control. The company’s focus on trust, transparency, and human collaboration may well define the next generation of decentralised finance tools.