AI trading agents have been connecting to broker platforms for just six months, yet regulatory frameworks remain notably absent. Bank of England Deputy Governor Sarah Breeden indicated in June that traditional human-in-the-loop oversight proves impractical for agentic trading speeds, suggesting regulators are instead considering kill switches and circuit breakers borrowed from algorithmic trading safeguards.
Major brokers including ThinkMarkets, Robinhood, IG Group, and eToro have already launched AI agent integrations, with infrastructure giants MetaQuotes and Spotware releasing their own versions. ThinkMarkets CEO Nauman Anees acknowledges no AI governance framework currently exists and enforcement would prove extremely difficult.
The EU’s ESMA issued supervisory guidance in February on AI-based algorithmic trading, though the EU AI Act lacks specific agent categories. The UK’s FCA published its Mills Review in July discussing agentic AI’s impact on financial services by 2030 but avoided direct trading regulation references.
Recent real-money AI trading competitions showed leading models from OpenAI to Alibaba mostly lost money and produced inconsistent decisions under identical conditions.
FXnCO Insight
Brokers deploying AI agents operate in a regulatory vacuum where liability for rogue agent behavior remains completely undefined, creating significant operational and legal risk.
Source: Finance Magnates