Current AI trading robots rely on a limited amount of historical data, which means that totally unknown market events, such as the 10/10 sell-offs or even last week’s severe sell-offs, will leave agent trading models short of the mark.
These historical data-driven AI models have never seen large liquidations in a single day and this would be “very unfamiliar” to them, Bitget CEO Gracy Chen said at a panel on agent trading robots at Consensus Hong Kong 2026. As such, human intervention is needed.
“As an exchange, we do not plan to build our own LLM [large language model]. But trading robots are a great thing,” Chen said. “Current AI robots are a bit like an intern: faster, cheaper, but they need a little supervision.”
However, later this will be more like a “full employee” and in 3 to 5 years AI will be able to replace many of us, Chen said.
These are sentiments heard regularly in the world of algorithmic trading when it comes to AI.
While the complex machine learning and LLM trading technology is improving rapidly, there are still many people who think that a human overlay is an essential part of the process, particularly in situations like the severe volatility that recently gripped crypto markets.
Joining Chen on the panel, Saad Naj, founder and CEO of agent trading startup PiP World, agreed that the technology is in its infancy and that comes with risks. But he noted that 90% of day traders and retailers lose money.
“As humans we are too emotional. We cannot compete with AI solutions,” said Naj.
