Artificial intelligence (AI) is becoming embedded across crypto trading, accelerating analysis, execution and optimization processes previously handled by people.
Investors and trading companies are being pushed to confront how much decision-making can be automated without diluting control, accountability or human judgment.
Even as some projects are reaching for more autonomous trading systems, most AI tools in crypto remain tightly constrained. Humans still define strategies, set risk limits and take responsibility for outcomes, as machines take on much of the bandwidth used for data-heavy tasks, such as research and monitoring.
Across crypto markets, the balance between automation and oversight is quietly reshaping trading workflows and beginning to redefine what human roles still matter.
“[AI is] replacing the 80% that nobody actually wants to do. The best researchers use AI to dramatically improve their work,” Ryan Li, co-founder and CEO of crypto research platform Surf AI, told Cointelegraph.
That change is already influencing how crypto trading firms operate, how junior roles are defined and where human judgment still sits in an increasingly automated market.
Crypto and trading job fears meet AI performance
Interest in using AI to boost efficiency in crypto accelerated in the last quarter of 2024 as AI agents emerged. Projects such as Virtuals Protocol drew attention for experiments involving AI-managed wallets and onchain activity.
Although AI agents remain overseen by humans, their growing potential has raised questions about whether traders will remain essential in future markets.
“From a technical point of view, autonomous trading is already possible. The question is not execution; it’s control, limits and accountability,” Igor Stadnyk, co-founder of AI trading platform True Trading, told Cointelegraph.
He added:
“But strategy selection and risk are still human decisions — you decide what to trade and how much risk to take. It’s your salary, after all.”
Related: Ethereum in 2026: Glamsterdam and Hegota forks, L1 scaling and more
Concerns about displacement extend beyond crypto. In traditional finance, researchers at Stanford University and Boston College tested an AI analyst using publicly available real-time data across thousands of US mutual fund portfolios between 1990 and 2020.
The AI-managed portfolios generated an average of $17.1 million more per fund per quarter than their human-managed counterparts. Ed deHaan, an accounting professor at Stanford who led the experiment, said he does not expect mass displacement of portfolio managers but warned that junior analyst roles could be at risk.
Describing candidates he evaluated but ultimately did not hire from his alma mater, Li said, “I’ve seen so many people with perfect scores from Berkeley, and they don’t know how to code. They don’t know how to write anything because they are entirely helped by AI.”
The remark was not a critique of the academic ability of modern students but an observation about how traditional hiring signals have weakened as AI tools take on work that once helped build foundational skills.
In crypto markets, decentralized perpetuals exchange Aster ran a separate experiment, pitting 100 human traders against 100 AI models during a period of market decline.
The competition ended with human traders down 32.21%. The AI models also finished in the red but preserved capital more effectively, posting a 4.48% loss.
AI trading is not algorithmic trading
Algorithmic systems now handle the vast majority of trade execution in major markets, replacing tasks once carried out by human traders.
Much of the concern around job displacement stems from treating AI trading as a continuation of algorithmic trading rather than a different class of systems altogether, Stadnyk said.
To put it simply, algorithmic trading is built around deterministic rules that execute predefined strategies when specific conditions are met, leaving little room for interpretation once those rules are set.
“With AI, you’re working under uncertainty, where data can be missing, noisy or even contradictory,” Stadnyk said. “AI is useful in those situations because it can still operate when information is incomplete and conditions are constantly changing.”
Related: Blockchains quietly prepare for quantum threat as Bitcoin debates timeline
AI can ingest and interpret news, social media and sentiment across regions and languages in real time, allowing traders to factor in narrative shifts and cultural context that are difficult to encode into fixed rules.
A similar pattern is visible at the network level, according to Nina Rong, executive director of growth at BNB Chain, where elevated trading activity has made shifts in trader behavior more visible.
“AI helps with gathering information for crypto folks and improves research efficiency, but only using information that’s already in the public domain,” Rong told Cointelegraph.
“It also gives non-programmers the ability to use programming as a tool. Domain experts who can use vibe coding to their advantage are in a uniquely strong position right now,” she added.
While AI is making traders more efficient, fears around job displacement continue to surface. In June, AI job replacement topped crypto social discussions, according to Santiment, a crypto research platform that uses AI to track market narratives.
Human judgment still matters in AI-driven crypto trading
AI has not removed humans from crypto, but it is already reshaping how work is distributed across the industry. Much of that shift is happening quietly, at the task level, particularly in research roles that once relied on teams of junior analysts and interns.
According to Li, those structures are already changing as AI absorbs routine research work that used to justify larger headcounts.
“Funds used to hire teams of researchers or interns,” he said. “Now they just have one really good researcher who can work with AI a lot better.”
But there are cases where AI systems have a higher degree of independence. In both crypto and traditional finance, autonomous models can be configured to manage wallets, rebalance portfolios and execute trades without constant human approval.
“I’m confident that major players are already doing this in some form, even if they’re not scaling it aggressively or promoting it publicly,” he added.
As execution becomes more automated, traders can focus on strategy and risk rather than manual mechanics. According to Stadnyk, the shift is happening faster than many expect.
“A year has passed since AI agents first gained traction on [X]. In crypto, that’s like 10 years in [aerospace] or 100 years in medicine because everything can be tested very quickly,” Stadnyk said.
Magazine: Chinese users turn to ‘U cards’ to get around crypto rules: Asia Express
Cointelegraph Features and Cointelegraph Magazine publish long-form journalism, analysis and narrative reporting produced by Cointelegraph’s in-house editorial team and selected external contributors with subject-matter expertise. All articles are edited and reviewed by Cointelegraph editors in line with our editorial standards. Contributions from external writers are commissioned for their experience, research or perspective and do not reflect the views of Cointelegraph as a company unless explicitly stated. Content published in Features and Magazine does not constitute financial, legal or investment advice. Readers should conduct their own research and consult qualified professionals where appropriate. Cointelegraph maintains full editorial independence. The selection, commissioning and publication of Features and Magazine content are not influenced by advertisers, partners or commercial relationships.
