Robinhood is rolling out AI agents that can analyze markets, develop trading strategies and execute trades for customers without requiring them to remain at their devices. While the technology is designed to automate trading, Robinhood says customers are responsible for the risks tied to orders placed by the agents.

The Nasdaq-listed company introduced “Robinhood Agents” Tuesday at its HOOD Summit in Houston. The feature is integrated into the Robinhood app, where users can select an AI agent and establish the limits under which it can operate.

Robinhood has more than 27 million funded accounts. The new system goes beyond the role of a traditional chatbot, which typically responds to user questions. Its AI agents can take actions, including buying and selling assets on a customer’s behalf based on the permissions and instructions provided.

The launch brings automated trading tools, historically used by hedge funds and quantitative investment firms, to a wider retail audience.

Robinhood first opened its platform to agentic trading in May, allowing technically skilled customers to connect their own AI agents to their accounts. More than 150,000 customers have opened agentic trading accounts since then, according to the company, and those agents now use Robinhood’s tools almost 30 million times each day.

With the latest rollout, customers can select and approve an AI agent without having to build one themselves.

“Agentic accounts come with trade approvals settings which you can configure to allow automated trade execution. With approvals on, your agent cannot place an order until you approve it. You can turn trade approvals off, and if you do, your agent can place orders without asking you to confirm each one,” Robinhood said in its announcement.

The brokerage is also preparing Loops, a feature that allows customers to turn a trading strategy into a recurring instruction for an AI agent. The system could be set to check the market each morning and trade when certain conditions are met, or continue operating overnight while the customer is asleep.

Robinhood said Loops is coming soon.

The development is part of a wider push to give AI agents greater control over financial activities. Meta’s Muse assistant can access users’ bank balances and investments, while x402, a payments protocol developed by Coinbase, allows agents to pay for services using stablecoins.

Customers remain responsible for AI-driven trades

Robinhood’s disclosures make clear that customers carry the financial risk from transactions executed through AI agents.

Customers “assume all risk for trades executed by AI agents and for any use of your data by third-party LLM providers,” the company said. Robinhood added that it “does not control, supervise, monitor, recommend, or audit agents.”

The warning is particularly important for Loops, which is designed to execute strategies repeatedly without requiring the user to approve every individual transaction.

Once activated, Loops “may place, modify, or cancel trades in your account automatically, without prompting you for approval on each transaction – including while you’re asleep, away from your device, or otherwise not monitoring the market.”

The feature will follow the customer’s strategy “exactly as configured, including during periods of market volatility.”

Robinhood does not guarantee Loops’ performance under any particular market conditions and says automated trading carries risks comparable to manual trading. Customers can disable the feature, but trades already executed by Loops will not automatically be reversed.

Concerns grow over autonomous market activity

The potential risks extend beyond individual accounts if autonomous trading agents become widely used.

Bank of England Deputy Governor Sarah Breeden warned in June that autonomous AI agents could “amplify volatility in stress” and potentially contribute to a “market meltdown.” She also said existing financial rules were not designed around agentic systems.

One concern is that large numbers of trading agents could react to the same market information in similar ways. If they simultaneously buy or sell in response to a news event, their activity could magnify an otherwise limited price movement.

Researchers from Wharton and the Hong Kong University of Science and Technology found in a simulated trading environment that AI agents could collude and manipulate prices to generate collective profits despite having no explicit communication channel.

The study also found that agents could achieve above-market profits without communication, prior agreement or intent, potentially complicating efforts to regulate autonomous trading systems.

The broader risks remain largely theoretical because agentic trading is still relatively new and adoption remains limited. The regulatory warnings and academic research concern AI-powered trading generally and do not establish that Robinhood’s agents behave in the same manner.

Nevertheless, the findings illustrate that autonomous trading introduces risks alongside the convenience of allowing software to make and execute investment decisions without continuous human oversight.