AI is narrowing the gap between investment ideas shared online and trades placed on crypto exchanges. MEXC CEO Vugar Usi says AI tools could give individual investors access to some quantitative trading techniques. But as AI agents gain more authority to act, investors also need to understand what they have authorized—and whether they can revoke it.
Social media shapes investment choices
Speaking at TOKEN2049, Usi said investment decisions increasingly sit alongside social content such as restaurant reviews and holiday photos. Seeing others succeed can prompt people to copy their choices. Investing can also become a way to express identity, whether by saying they backed Amazon early or spotted an opportunity first.
A study published by the FINRA Investor Education Foundation in April, using data from its 2024 investor survey, found that 59% of investors who use social media said social activity influences their motivation to invest. The figure was 11% among those who do not use social media. The study also found a gap between social media users’ confidence and their tested investment knowledge. Social content can lower the barrier to engaging with markets, but it does not necessarily help investors judge whether a recommendation is sound.
Usi said interest in artificial intelligence companies has also helped drive trading in related assets on MEXC. Popular themes can bring opportunities to investors’ attention, but a theme’s visibility alone says nothing about whether an asset is fairly priced.
MEXC connects user-built AI agents to its exchange
On September 28, 2026, MEXC introduced a tool that lets users connect their own AI agents to the exchange to prepare and execute trades, with user confirmation still required for key actions. Usi said AI could help individual investors draw on algorithmic approaches long used by quantitative trading firms. But professional firms also rely on tested strategies and dependable data; access to an AI assistant does not mean an individual can replicate a quant fund’s performance.
A survey published by MEXC on October 15 covered 3,708 crypto traders. It found that 84.8% of respondents had used AI for research or trading. Among those who had, 65.9% said they acted on AI recommendations at least sometimes. These figures reflect MEXC’s survey sample. Respondents’ reports of improved performance represent their personal impressions, not independently verified investment returns.
Usi said a possible next step is for autonomous agents to adjust portfolios and pursue opportunities on users’ behalf. As agents take on more tasks, investors will need to assess how they make decisions and whether they can understand the actions taken.
Longer trading hours add to demand for automation
Usi noted that exchanges operate continuously, while people need to rest. Nasdaq plans to introduce a five-day-a-week, 23-hour trading schedule from December 6, giving investors more time to respond to markets.
Usi envisaged AI agents managing positions while investors are asleep and acting before a position approaches a dangerous level. He suggested such tools could reduce reliance on forced liquidations and stop-losses, but the interview provided no evidence that AI can prevent leveraged positions from being liquidated. An agent’s ability to run around the clock does not ensure its decisions are correct.
Longer trading hours also give investors more time to act—including impulsively. Automation may help users follow plans they have set in advance, provided the system adheres to the limits they specify.
A confirmation button is not enough
The current process Usi described retains a human checkpoint: AI proposes a trade, and the user confirms it before execution. But clicking confirm is no substitute for understanding the trade. A confidently worded recommendation may still rest on weak foundations. Usi said financial firms should approach AI cautiously, taking its risks and limitations seriously rather than readily removing safeguards.
Guidance on the responsible use of AI published by the Financial Stability Board on October 15 says human oversight should be proportionate to a system’s autonomy and risk. Effective supervision involves more than approving actions; it also requires the ability to intervene when necessary.
The MEXC survey also found that 54% of respondents would allow AI to trade no more than 20% of their account assets if it could not transfer or withdraw funds. Just 6.5% were willing to let AI trade more than half of their assets. The responses point to limits many respondents would place on AI trading access, but the survey should not be taken as representative of all crypto investors.
Usi stressed the importance of the relationship between platforms and their customers. For platforms that provide investment information, trading tools and execution, the scope of user authorization is particularly important. As AI agents take part in more decisions, investors need to know which permissions they have granted—and whether they can withdraw them when a strategy no longer suits them.