Recently, after the popularity of OpenClaw, some investors have started applying the tool to the stock market for reviewing past trades, selecting stocks, and using it as a "trading assistant." But is relying on OpenClaw for stock trading actually reliable?
One user reported significant losses after following OpenClaw’s recommendations. Between March 13 and March 17, the tool suggested three stocks, each of which resulted in a loss. Over three trading days, the user lost a total of 7,000 yuan. On March 17 alone, OpenClaw advised cutting losses on a power equipment stock, which had declined by 6.33%. Minutes later, it recommended buying a photovoltaic concept stock, which fell 3.27% by the market close.
The user currently expects OpenClaw to provide a daily trading plan before the market opens, update holdings every 20 minutes during trading hours, and summarize the day’s performance afterward to help the tool "evolve." However, the user found that OpenClaw’s stock-picking logic largely relies on the previous day’s top trader rankings. Another issue is the tool’s limited memory—repeated instructions are often forgotten.
For another user, OpenClaw serves as a scheduled assistant that uses Python and financial APIs to automatically review the market based on preset logic. It scores stocks and generates reports. This user noted that, while the tool improves efficiency, its stock recommendations have not been profitable—out of 12 recommended stocks, only two saw minor gains while the rest fell.
According to industry experts, AI tools like OpenClaw can enhance efficiency but do not improve trading success rates. Large language models have inherent limitations in processing large-scale data and cannot perform full-scope analysis. Their real value lies in speeding up information processing, such as sorting through company data or building information networks. However, they remain an extension of the user’s own capabilities, not a revolutionary trading tool.
Experts warn that investors should avoid overestimating AI’s capabilities. Without a solid trading strategy, relying solely on AI for profit is unlikely to succeed. Unlike high-frequency quantitative trading, which operates in milliseconds, AI recommendations often come with delays, reducing their effectiveness in fast-moving markets.
From a technical standpoint, the rise of AI-assisted trading stems from improvements in generative AI and the shift from passive interaction to active task execution. OpenClaw and similar tools allow users to create basic investment analysis workflows using natural language. Yet, whenever new technology enters the financial market, there is a tendency to overstate its capabilities. While AI can improve research efficiency, investing remains a highly uncertain process.
Risks also accompany the use of such tools. Regulatory gaps, data security issues, and unclear accountability are major concerns. If an AI tool with full system access causes data leaks, executes erroneous trades, or even engages in unauthorized activities, it is unclear who would be held responsible—the user, developer, model provider, or the AI itself.
Moreover, many AI-based stock-picking platforms are not licensed to provide investment advice. If losses occur, assigning liability becomes complicated. Users are advised not to treat AI as a substitute for decision-making. Instead, a "machine alerts, human decides" approach is recommended—using AI for data processing and preliminary analysis, while verifying key information through authoritative sources.