Beyond the AI hype: Why the Future of Trading Depends on Responsible Innovation and not Marketing Claims

Interview

Beyond the AI hype: Why the future of trading depends on responsible innovation and not marketing claims

AI has quickly become the defining narrative of global innovation. Across industries, companies are repositioning themselves through the language of intelligent assistants, agents, automated insight, and predictive systems. The fintech and trading sectors have followed suit: trading platforms increasingly showcase AI-driven features, promising personalised analysis or accelerated decision-making.

However, this rapid adoption raises a deeper question, one that the industry has not yet clearly answered: Is AI truly transforming the trader’s experience, or has it become the latest technology deployed faster than its purpose is understood?

At the same time, retail traders themselves have evolved. As highlighted during a recent panel discussion, many now use external LLMs (Large Language Models such as chatGPT) to gather context, interpret news, and generate trade ideas long before they open their trading platform. The technological landscape has shifted—but not always in the ways the industry expected.

We spoke with Milica Nikolic, Exness trading product operations leader, about what innovation should look like in the new environment, and why responsible progress may require a more thoughtful approach than simply adding AI to the front end of trading.

The big AI question:

Over the past two years, AI has become part of the trading environment in a very visible way. From your perspective, how has AI transformed the retail trading space so far?

We all see it: AI has become part of our everyday rhythm. We use it to summarize long articles, plan our schedules, clarify unfamiliar concepts, or simply speed up tasks that used to take much longer. In this sense, trading is no exception; the same tools that help us organize information in daily life are now shaping how traders prepare for the markets.

What AI has really introduced into retail trading is a new level of accessibility. Traders can gather context, interpret events, and review market history in seconds, using systems that condense large volumes of information into clear summaries. Pre-trade preparation has become faster, more structured, and in many cases, more consistent.

However, the fundamental nature of trading remains unchanged. AI can support research, highlight themes, or explain recent movements, yet interpretation, risk assessment, and decision-making still rely on the trader. AI has expanded access to information, but it hasn’t replaced the skill required to act on it.

The “AI-powered” label on every single solution:

What does an “AI-powered retail trader” look like today? How is their behaviour different from what you observed in 2023–2024?

The clearest change is where the trading process now begins. A few years ago, most retail traders started inside their trading platform: opening charts, checking technical indicators, reviewing calendars, and building their outlook from there. Today, many begin much earlier and in a completely different environment.

An “AI-powered” trader often forms their initial view before they even log in to the trading platform. They use AI tools to summarise market sentiment, interpret recent moves, or understand what events might influence the day ahead. By the time they reach the trading terminal, they already have a framework in mind—a narrative that guides what they look for and how quickly they act.

This creates a more prepared trader, as well as a faster one. They process information more efficiently, but the core behaviours: discipline, patience, and risk awareness still differ from person to person. AI doesn’t standardise the decisions that follow; it simply accelerates the research phase.

As AI features appear across trading platforms, Exness has taken a deliberately measured path. Is this a matter of waiting until the technology truly adds value, an approach often associated with Apple, or is it part of a longer-term AI strategy?

Our philosophy is very close to that: we introduce technology when it genuinely improves the trading experience, not simply because it is fashionable. The starting point is clarity around the broker’s role: what we should and shouldn’t do. AI can organise information, summarise context, and help traders understand the environment more efficiently. But the moment it begins suggesting actions or framing decisions, it risks crossing into advisory territory. That boundary matters.

So, when it comes to client-facing AI, we are guided by whether the technology can genuinely enhance clarity without shaping a trader’s choices. If it doesn’t meet that standard, it isn’t ready, or it’s not something that we are looking into, regardless of how advanced or popular the underlying models may be. In that sense, we prefer to release technology only when its purpose is clear, its value demonstrable, and the experience frictionless.

At the same time, AI already plays a significant role in our product. We use models that have a tangible impact on the client experience our traders receive. That’s where we see the greatest responsibility for a broker: using technology to make a client’s trading experience better, not to influence a trader’s judgment.

Client-facing AI may have a place in the future, but only if it supports autonomy, rather than substituting it. This underscores a larger point: any meaningful use of AI must be grounded in robust safeguards.

What safeguards do you believe should become standard in the future?

Responsible AI must be built on strong guardrails at the model and data level. This includes favouring domain-specific models trained on vetted, factual sources rather than broad internet data. Outputs should be grounded in referenced, structured data, with any assumptions explicitly stated, avoiding reliance on retained memory or generalized knowledge.

AI systems should retrieve and reference source data at the time of each response, rather than inferring context or relying on persistent memory. This ensures answers are based on current information and can be traced directly back to underlying documents, prices, or policies.

Outputs should clearly distinguish between factual statements, interpretation, and explanation. This separation improves predictability, enables auditing, and makes AI systems suitable for regulated environments.

Clear human ownership is essential across training, deployment, and monitoring. Human oversight ensures outputs remain factual, compliant, and aligned with internal policies, even as models become more capable.

Robust model governance requires pre-deployment evaluation and continuous testing throughout the system’s lifecycle. This includes monitoring for data drift, outdated information, hallucinations, bias, and deviations from expected behavior.

Together, these safeguards establish not just accuracy, but a framework in which AI behaves consistently, responsibly, and in line with the expectations of a financial institution.

As AI becomes a larger part of the market’s infrastructure, what responsibilities do brokers have in ensuring that innovation remains transparent and aligned with traders’ interests?

Brokers have a responsibility to ensure that innovation does not outpace the safeguards that protect traders. As AI becomes more integrated into the systems that support pricing, execution, or market intelligence, transparency becomes essential, not only in how these technologies operate, but in how their outputs should be interpreted.

The first responsibility is clarity of role. Brokers must avoid using AI in ways that could influence a trader’s decision or resemble advice. Innovation should support autonomy, not override it. Then, brokers have a responsibility to ensure that the pursuit of novelty does not dilute trust. Not every technological advance needs to be visible to the client; in many cases, the most meaningful innovations are those that strengthen the foundation, execution quality, pricing coherence, and platform resilience, without changing the decision-making process itself.

Final thoughts

AI is reshaping the way traders gather information and form their views, but it hasn’t changed the core principles that define responsible participation in the market. Traders still depend on clarity, stability, and autonomy, and the systems around them must reinforce those conditions rather than compete with them.

As AI becomes more integrated into the market’s infrastructure, the real measure of progress will not be how visibly it appears in front-end experiences, but how reliably it supports the environment in which decisions and trades are made. The most meaningful innovations are often those that remain largely unseen: models that enhance execution quality, systems that maintain coherent pricing under pressure, and safeguards that ensure technology behaves predictably even during volatile moments.

In that sense, the next stage of AI in trading is less about reinvention and more about responsibility: ensuring that the tools shaping the market evolve in ways that protect its integrity and support the people who participate in it.