70% US Traders Want AI Portfolios, But Demand Revocable Control

Woofun AI reports that a survey of 1,400 US crypto traders conducted by OKX indicates 70% are comfortable with AI managing portfolios, either fully or within user-defined limits. This figure does not imply a blanket acceptance of unrestricted algorithmic access to savings. Instead, it reflects a nuanced preference where traders are willing to allow automated trading only after strictly defining available capital, supported assets, and acceptable risk parameters. The distinction between full control and limited control is critical, as it underscores that while interest in automated execution is high, the desire for unsupervised trading remains low. Crypto platforms are already building the infrastructure to move AI from analysis into execution, but the market’s appetite is conditioned on maintaining oversight.

The infrastructure for agent-enabled trading is being developed by multiple major players, each with distinct approaches. OKX and eToro have launched tools that support agent-enabled trading, allowing for varying degrees of automation.

Meanwhile, Robinhood has taken a different path by launching agentic accounts for traditional securities, with a planned expansion into crypto. This competitive landscape highlights a shift in how trading platforms are integrating AI, moving beyond simple advisory tools to systems capable of executing trades. The entry of traditional finance giants like Robinhood into this space signals a broader industry trend toward agentic trading, where AI agents can operate within defined boundaries. This expansion is not limited to crypto; it encompasses equities and options, suggesting a universal application of these technologies across asset classes.

AI portfolio management encompasses a spectrum of autonomy levels, ranging from minimal intervention to full independence. At the lowest level, a chatbot may summarize market news or explain position movements, leaving all decision-making and order placement to the user. An approval-based agent takes a step further by preparing trades, calculating position sizes, and waiting for user confirmation before execution. A limited autonomous agent can execute trades automatically but is constrained by rules such as a maximum budget, a list of permitted assets, or a ban on leverage. Full autonomy grants the system the freedom to select and execute strategies without transaction-level approval. The OKX survey’s headline figure of 70% comfort includes both limited autonomous and full autonomy categories, indicating that while many are open to automation, few are ready to relinquish all control.

Demographic divides significantly influence trust in AI autonomy, with younger generations showing greater willingness to adopt unsupervised trading. Among Gen Z respondents, 38% said they would allow an AI to operate without direct supervision, compared to 37% of Millennials. In stark contrast, only 11% of Boomers gave the same answer. This generational gap extends to trust in AI-generated recommendations versus human advice. A quarter of Gen Z and Millennial respondents trusted AI recommendations more than human advisers, while the share among Gen X and Boomer respondents was roughly half as large. These differences highlight that age is a key variable in the adoption of AI trading tools, with younger traders more inclined to embrace automation and older traders preferring human oversight.

The shift toward automated execution is part of a broader change in how traders gather and process information. Fifty-one percent of respondents use AI for research or trading several times a week, and 77% had used a general-purpose chatbot to investigate a crypto position in the previous three months. Research is a low-risk entry point, allowing traders to compare AI-generated answers with price charts, company announcements, or regulatory filings before acting.

However, connecting the same system to an exchange account changes the stakes significantly. A misunderstood instruction or incorrect parameter can result in a real position within seconds, turning a research tool into an execution engine. This transition from analysis to action introduces new risks, as the consequences of errors become immediate and financial.

The distinction between AI agents and traditional trading bots lies in their flexibility and adaptability. Traditional trading bots follow rules written in advance, executing trades based on predefined conditions. In contrast, an AI agent can interpret broader instructions, decide which tools to use, and adapt its response as new information becomes available. This flexibility makes AI agents easier to use but also creates more room for unexpected behavior. While traditional bots are predictable within their rule sets, AI agents can deviate from expected patterns, potentially leading to unintended outcomes. This adaptability is both a strength and a risk, as it allows for dynamic responses to market changes but also increases the potential for errors.

OKX has released the Agent Trade Kit, a tool that allows compatible AI agents to interact with exchange accounts. This kit supports spot, futures, options, and advanced order execution through natural-language instructions. Depending on the permissions granted, an agent can inspect balances, monitor positions, place or amend orders, and establish stop-loss or take-profit levels. It can also run automated strategies such as dollar-cost averaging or grid strategies.

To mitigate risks, OKX supports simulated trading and read-only access, allowing users to test agents without risking live funds. The exchange recommends using a separate sub-account and limiting it to the amount intended for the strategy. It also warns that models may misunderstand instructions, rely on outdated information, or execute during periods of poor liquidity and high slippage. These features emphasize the importance of controlled access and risk management in AI trading.

Woofun AI data shows, Robinhood and eToro have also launched products that reflect this trend toward agentic trading. Robinhood launched Agentic Trading accounts in May 2026, initially supporting equities and later adding options. Customers can connect a third-party AI model to an account reserved for agent activity, with the system accessing only the capital placed inside that account. Users can monitor trades, follow profit and loss, receive activity notifications, and disconnect the agent.

Robinhood announced a planned rollout of Agentic Accounts for crypto trading, allowing eligible US customers to connect an AI model to its crypto data and trading tools. eToro introduced Agent Portfolios in March 2026, allowing investors to create a dedicated portfolio, assign a budget, and connect an AI through an API key restricted to that portfolio. The agent can inspect balances and open or close positions within the assigned funds. Both platforms emphasize limited access, ensuring that AI agents operate within defined boundaries.

Trust in AI agents is heavily influenced by the ability to maintain control and receive timely information. When asked what would make them trust an AI agent with payments, respondents prioritized real-time notifications and the ability to revoke permissions immediately. This answer was selected more than twice as often as any alternative and remained popular across age groups. The emphasis on revocable access suggests that users are willing to accept automated decisions as long as they can quickly intervene if necessary. This preference for control extends beyond investment accounts, as AI agents begin to pay for services and complete transactions independently. The ability to revoke permissions is a critical factor in building trust, as it ensures that users retain ultimate authority over their assets.

The performance of AI-generated trading recommendations varies, with 55% of respondents describing the outcome as successful, 41% calling it mixed, and 4% saying it had backfired. These results are based on personal assessments rather than verified portfolio returns and were not compared with Bitcoin, a market index, or a passive strategy. Different respondents may define a successful recommendation differently, making it difficult to draw definitive conclusions about AI performance. The survey does not disclose how participants were recruited, when the fieldwork took place, whether the sample was weighted, or what margin of error applies. Therefore, while the findings indicate interest in AI trading tools, they do not prove that AI-managed portfolios outperform human traders or established automated strategies.

Regulatory scrutiny is increasing, with the US Securities and Exchange Commission (SEC) including automated investment tools, AI technologies, and trading algorithms in its 2026 examination priorities. The SEC may examine whether statements about AI capabilities are accurate, whether systems operate consistently with disclosures, and whether automated recommendations remain appropriate for an investor’s profile. For supervised financial firms, using a third-party model does not remove the need for controls. Regulators may still examine how products are described, what agents are allowed to do, and whether customers understand the authority they have granted. This regulatory focus underscores the importance of transparency and accountability in AI trading, ensuring that users are fully informed about the risks and capabilities of these tools.

The demand for AI tools that can act rather than simply advise is clear, as evidenced by the launches from OKX and eToro. Robinhood’s planned crypto rollout could bring this model to a wider retail audience, further normalizing the use of AI in trading.

However, a long and comparable performance record across different market conditions is still missing. Crypto trades continuously and produces large amounts of real-time data, making it a natural testing ground for AI agents. Yet, these same characteristics allow a flawed strategy to keep operating while the account holder is offline. Early products are therefore being built around limited autonomy rather than unlimited access, ensuring that investors retain control over their capital. The agent may research, monitor, and execute, but the investor decides how much capital it can reach and how quickly that access can be removed. Whether these tools become widely trusted may depend less on how often an AI finds the right trade and more on what happens when it gets one wrong.