Nickel Digital leader states AI is not a solution during challenging market conditions, but it can provide assistance.

At times, human involvement and a traditional approach are necessary, remarked Anatoly Crachilov, founding partner and CEO of Nickel Digital Asset Management.

Anatoly Crachilov, CEO of Nickel Digital (Nickel Digital modified by CoinDesk)

Key points:

• A recent study conducted by Nickel Digital Asset Management revealed that nearly all trading firms reported AI has become integral to their main investment strategies.
• However, there are instances where entirely automated trading may not suffice, necessitating human judgment.
• In spite of the current downturn in cryptocurrency markets, London-based Nickel, managing a multimanager platform with allocations to over 80 teams, maintains a positive outlook for the year.

As market conditions become challenging, like the crypto situation at the end of January, investment firms require all available resources to make swift and accurate decisions. It is therefore not surprising that many are resorting to AI as a cutting-edge tool to analyze and propose methods for reducing losses and potentially generating profits.

According to recent research from Nickel Digital Asset Management, almost all (96%) executives from a surveyed group of trading firms managing approximately $14 trillion in assets indicated that AI already plays a significant role in their core investment processes. Nonetheless, more is needed, as a human touch remains essential, stated Anatoly Crachilov, founding partner and CEO of the company.

STORY CONTINUES BELOWDon’t miss another story.Subscribe to the Crypto Daybook Americas Newsletter today. See all newslettersSign me upBy signing up, you will receive emails about CoinDesk products and you agree to our terms & conditions and privacy policy.

AI is altering quantitative trading much like it is transforming nearly every other sector and human activity. Beyond the large language models (LLMs) that seem to have infiltrated many aspects of daily life, there are also machine learning and predictive AI techniques that analyze past data to predict future trends. However, these techniques often struggle to identify incorrect information that could result in misleading conclusions and poor decision-making.

“It’s a very tough market. AI will not save you; it’s not a savior,” Crachilov stated in an interview.

Despite the decline in cryptocurrency values that swept through the market at the end of last month, London-based Nickel, which operates a multimanager platform allocating to over 80 teams, remains optimistic for the year ahead. “Perhaps an achievement in its own right,” Crachilov noted.

The intersection of cryptocurrency trading and AI is advancing significantly in areas such as risk management. Although AI may still find it challenging to outperform high-speed trading bots focused on the latest low-liquidity crypto tokens, there exists potential in areas where sentiment and data-driven models can effectively manage risk.

Each manager affiliated with Nickel operates within a clearly defined risk framework that encompasses maximum drawdown limits during periods of heightened volatility. Occasionally, human intervention is necessary, along with a “traditional” approach, Crachilov described, rather than solely depending on data-driven, machine-learned automation.

“If the market experiences distress, as it has on several occasions recently, there are times when you must enforce discipline and halt those managers who exceed [max drawdown] limits, regardless of whether AI is steering their strategy,” Crachilov remarked. “Ultimately, there is a firm limit on how much distress we would permit in the portfolio.”

Discussions regarding the extent of human participation in AI-led trading strategies, or how a human override should be activated, were considered too intricate and detailed for Nickel’s relatively broad survey of managers, Crachilov explained.

He added that Nickel functions as “a military-style operation,” where a continuous data flow collects over 100 million data points from the underlying book every 24 hours. “While this aspect is very well informed, it still necessitates human involvement. We’re still in contact with managers, even in the middle of the night,” Crachilov stated.

The natural progression towards complete automation must still consider the potential for flawed or incomplete data feeds from sources like cryptocurrency exchanges, according to Crachilov.

For instance, a human would recognize that data showing a particular position is down 100% likely indicates an issue with a data feed. However, an automated AI system might rigidly enforce a limit when it wasn’t warranted.

“A human overlay is essential. The entire cryptocurrency ecosystem remains quite fragile. Some exchanges may experience downtimes of 15 minutes, or present incorrect data, or produce segments of faulty data, which could inadvertently lead the system to shut down some managers without justification,” Crachilov explained.

Ultimately, it boils down to the firm’s risk management philosophy, which aims to eliminate any single point of failure throughout the process, according to Charles Adams, Nickel’s head of investor relations.

“If there were one autonomous agent monitoring the entire portfolio, and something were to go awry, the risks could be tremendously severe,” he noted. “The primary objective is to maintain a highly diversified fund distributed among more than 80 managers today across hundreds, if not thousands, of sub-accounts on exchanges, and it is crucial for us to eliminate that single point of failure.”