Federico Variola (페데리코 바리올라), CEO of cryptocurrency exchange Phemex, said artificial intelligence has been broadly negative for the crypto industry.
Cointelegraph reported on Sept. 16 that he said AI is siphoning money from the industry and boosting attackers' capabilities, while also driving up cybersecurity costs for small teams and increasing pressure toward centralisation in the crypto sector.
Variola appeared on Cointelegraph's "Chain Reaction" and said it is hard to be optimistic about AI in crypto. He said liquidity has shifted significantly to the AI industry and that the use of AI for social engineering attacks and vulnerability discovery has made actors who exploit protocols stronger.
He said the changes could make it harder for small teams developing protocols to operate without massive cybersecurity budgets. He also said AI threats could reduce the appeal of self-custody and decentralised finance, because growing threats such as device hacking and social engineering attacks increase the burden individual users must bear.
The assessment came after Phemex announced an AI transformation plan in February. At the time, Phemex said it would introduce AI across product development and internal operations, but the latest remarks focused on the crypto industry overall rather than the company's strategy.
Recent hacking cases also align with those concerns. In July, about $116 million worth of bitcoin was drained from more than 5,200 addresses affected by a flaw in the Coldcard hardware wallet. It has been widely suggested the flaw may have been found through malicious use of AI. At the time, Coinkite CEO Rodolfo Novak (로돌포 노박) warned that AI-assisted code reviews can find bugs faster than skilled industry experts.
Potential defensive uses of AI have also been raised. CertiK chief blockchain investigator Natalie Newson (나탈리 뉴슨) said in April that while AI is making attacks more sophisticated, it could also become one of the most powerful defensive tools.
Variola said he recognises the practical usefulness of AI agents. He said AI agents can help improve an investor's portfolio construction or trading judgement, but they will not be able to make final trading decisions on the investor's behalf.