18 Nov 2026, Wed

Animoca Brands Chairman Yat Siu Warns Crypto Traders Over the Hidden Risks of AI Agent Control

As artificial intelligence agents rapidly evolve to navigate the 24-hour cryptocurrency markets on behalf of human users, industry leaders are beginning to weigh the profound conveniences of autonomous technology against its underlying structural risks. Speaking on the sidelines of the Token2049 conference in Singapore, Animoca Brands Chairman Yat Siu cautioned that while crypto traders can successfully put AI agents to work around the clock, users simultaneously risk building their entire financial and personal routines around services that they ultimately do not control.

The promise of autonomous AI agents in the digital asset sector is compelling. Unlike traditional static chatbots that merely answer direct prompts, modern AI agents can actively search for new information, analyze market trends, and alert their users to critical developments that they might otherwise miss during sleep or busy schedules.

To illustrate the operational difference, Siu pointed to conference planning. While a standard chatbot can supply a predefined list of upcoming events when explicitly asked, an autonomous agent operates proactively. It scans for newly announced schedules, tracks relevant industry developments, and suggests attendance when a worthy event pops up. In the context of digital assets, this proactive capability translates to round-the-clock market surveillance. Traders can leverage agents to monitor price action, liquidity shifts, and breaking news without needing to remain glued to their trading terminals.

However, despite their immense potential, the technology still suffers from distinct teething problems. Siu shared a personal anecdote regarding his own experimental AI agent, which became so intensely focused on crypto trading and investment management that it repeatedly attempted to interpret completely unrelated lifestyle questions through a cryptocurrency lens. According to Siu, the practical solution to this behavioral overreach is to abandon the concept of a single, all-encompassing assistant in favor of running multiple, highly specialized agents.

The blending of personal and financial tasks poses a significant operational hazard. If a single agent is tasked with handling a user’s personal calendar, family schedule, travel arrangements, and on top of that provides high-stakes advice on crypto investments, the arrangement quickly becomes confusing not just for the artificial intelligence, but for the human user as well.

The Questions That Set Trading Strategies Apart

Beyond operational confusion, deploying an AI agent does not inherently grant a trader a competitive market edge. Because foundational large language models are widely accessible, a thousand different traders asking an artificial intelligence about the exact same potential trade could theoretically receive roughly the same standardized response.

Agents can trade 24/7 — but Yat Siu warns you could lose access at any time

Regulators have increasingly taken note of this homogeneity, recognizing that widespread reliance on AI-driven traders could lead market participants to pile into identical positions simultaneously. The Financial Policy Committee of the Bank of England issued a formal warning highlighting that the widespread use of similar algorithmic and AI models could cause investors to take correlated positions during periods of financial stress, thereby dangerously amplifying market shocks.

Yet, this standardization is not inescapable. AI responses diverge significantly when users introduce follow-up questions deeply rooted in their own unique knowledge, personal constraints, and subjective concerns. As a user interacts with an agent over time, their personal tastes, ideas, and strategic philosophies begin to shape the system’s output. These ongoing, individualized inputs help mold generic queries into personalized crypto trading strategies that reflect a distinct human perspective.

AI Agents Receive Wallets and Trading Tools

The integration of artificial intelligence into the digital asset ecosystem has moved far beyond theoretical discussions. Major cryptocurrency firms are already equipping AI agents with the direct financial tools necessary to handle capital alongside information.

In February, Coinbase introduced Agentic Wallets, a feature designed to allow autonomous agents to spend and trade cryptocurrency independently within strict limits pre-determined by human users. Shortly thereafter, Binance launched AI Agent Skills in March, granting artificial intelligence systems direct access to live market data alongside the technical ability to place and manage spot orders.

These advanced tools allow retail and institutional users to delegate a larger share of the heavy lifting required to follow volatile global markets around the clock. By granting agents the capacity to execute transactions within predefined boundaries, traders can potentially capture lucrative market opportunities during the late-night hours without sacrificing their sleep or personal well-being.

Despite these advanced capabilities, human input remains vital—not necessarily in the form of a constant human-in-the-loop approval mechanism for every single transaction, but through the overarching ideas, queries, and strategic directions that guide the agent’s behavior. As Siu emphasizes, the artificial intelligence learns a user’s unique preferences over time through iterative direction, where the human ultimately dictates the high-level focus by pointing the system toward what truly matters.

Agents can trade 24/7 — but Yat Siu warns you could lose access at any time

The Privacy Cost of an AI That Knows You

The convenience of an assistant that anticipates your every need comes with a steep privacy cost. Siu suggests that users ought to treat an advanced AI assistant with the same cautious distance they would apply to an external third-party financial adviser. Before handing over sensitive documents or personal logs, users must carefully consider what specific information the agent genuinely needs to formulate an accurate answer and what sensitive details can be safely left out.

The accumulation of personal data creates deep vulnerabilities. A user might spend months carefully training an agent, teaching it which specific markets to follow, what information sources to digest, and what risk parameters to observe before executing a trade. Over time, the agent retains these conversational contexts to save the user from repeatedly explaining the same core preferences.

However, this accumulated intelligence resides within a centralized framework. The primary danger is that the corporate entity running the underlying service maintains absolute control over user access. A sudden account suspension, a corporate policy shift, or a technical platform outage could interrupt everything a trader is doing all at once, severing their connection to their customized tools instantaneously.

Siu asks users to imagine an assistant that has steadily learned intimate details about their financial habits, risk tolerance, and personal life over a period of years, only to vanish entirely on any given morning due to platform restrictions. To mitigate this systemic risk, users must demand fundamental data portability rights. At a bare minimum, individuals should retain the legal and technical right to download their accumulated AI profile and interaction data so that it can be securely transferred elsewhere, allowing another agent to take over seamlessly without forcing the trader to start from scratch.

By Basiran

Leave a Reply

Your email address will not be published. Required fields are marked *