Five key takeaways from FIA L&C 2026
Douglas Moffat
•
Senior Vice President, Americas
May 1, 2026
The FIA’s Law and Compliance conference is a key event in the regulatory calendar. With a strong focus on legal, regulatory and enforcement developments, the event brings together regulators and regulated firms across over 30 sessions to discuss the key issues impacting compliance teams today.
As ever, our team was on the ground, attending these talks and speaking directly with delegates. Here are our five top takeaways from the event.
"What we're finding is that the human touch on these investigations is not going away. There are great tools that allow you to execute a little faster, but all of this still requires a pretty strong human touch.”
Predictably, the accelerating use of AI was a strong point of focus across the entire event. With new technologies being introduced to the market at an unprecedented rate, there has been uncertainty about what impact it will have on compliance strategies and the role that AI will play.
However, across a number of sessions focusing on this topic, the general consensus was that AI should be regarded as a supplement – not a replacement – for compliance and surveillance functions. AI technologies can dramatically improve the speed, scale, and efficiency of investigations into potentially suspicious trading activity, but they cannot replace the human judgement necessary to analyze and report on this activity.
By framing the use of AI as a tool as opposed to a decision-maker, compliance teams will benefit from improved efficiency without the risk of poor explainability, which regulators hate. And on that note…
“On surveillance, I think explainability is the key thing, right? Understanding the information that's coming out of your system [...] and having the necessary tools and documentation steps to support your conclusions and analysis.”
It’s no secret that regulators are currently pushing firms towards the need for regulatory explainability. With the increased use of AI-powered tools for trade surveillance, many firms are now at risk of losing touch with the “why” behind their strategies.
To be truly compliant, you must be able to show:
In other words, it’s no longer sufficient to simply have a surveillance system in place – firms must be able to explain why it has been calibrated in the way it has and demonstrate that it allows them to correctly identify instances of market manipulation.
If firms fall into the trap of “black-box AI” - allowing AI-powered systems to make decisions unchecked on their behalf - they will naturally lose the ability to explain why certain alert thresholds have been set and will run the risk of non-compliance. The only way of avoiding this is to ensure proper human oversight – strong controls around data-quality and system calibration allow firms to leverage new AI technologies without risking a loss of explainability.
“AI over time might develop and help that be more efficient. But [...] there's still a lot of manual alerts. What we do quite frequently is continuously look at our thresholds and our orders, we back test and we say ‘can we tweak this or tweak that to reduce the noise but not miss anything?’ - It's a constant struggle.”
The noise generated by high volumes of false positive (and false negative) alerts remains one of the biggest challenges for surveillance teams. Poorly calibrated or overly static alert thresholds can lead to high volumes of misleading alerts being flagged in your surveillance system, which makes it much more difficult to isolate genuinely suspicious trading activity.
Depending on their trading volume, some firms will face the challenge of managing thousands of alerts every day. However, if there is a lack of suitably trained staff or insufficient governance processes to properly review them, this volume will create a state of alert fatigue in compliance teams. The risk here is that firms will attempt to solve this with an over-reliance on AI-powered investigation to clear alerts as quickly as possible. Without proper human oversight and intervention, this approach means that they risk falling into the trap of poor explainability as discussed above.
On a more positive note, the view of Thursday’s “Trade Surveillance: Man vs. Machine” panel was that there was reason to be optimistic about the future. Big strides are currently being made by regulatory technology providers in their mission to help firms reduce false positives, with many seeing that AI-assisted surveillance tools are already greatly enhancing the efficiency of compliance teams.
At eflow, reducing false positives is one of the key objectives behind our trade surveillance technology. Dynamic and highly configurable threshold calibration, combined with explainable AI investigations, allows firms to cut through the noise of false positives and get to genuine risk quicker.
“Many of us remember having to rely on off-the-shelf systems for catching spoofing when it was just in a single market. And now we're talking about cross market surveillance for prediction markets.”
As markets become more interconnected and as new asset types such as prediction markets and digital assets mature, the most serious integrity risks are increasingly cross‑product and cross‑market in nature.
Modern manipulation strategies exploit differences between markets: variances in liquidity, trading hours, reference pricing and transparency provide avenues which can be exploited by bad actors. In isolation, these individual trading behaviors may seem legitimate, but it is only once multiple actions across markets are stitched together that the wider context can be understood and the abusive trading identified.
This issue has been exacerbated by the rapid adoption of prediction markets. Their simplicity, binary pay‑offs and information‑driven pricing make them attractive both for legitimate participants and for sophisticated manipulation strategies that are executed across venues and products.
Prediction markets are not necessarily creating new forms of abuse but rather revealing weaknesses in legacy surveillance architectures. The weaknesses exposed by prediction markets are equally applicable to other markets, including more traditional financial assets.
The solution? ‘Data fusion’. By integrating data from multiple sources holistically in a single system, compliance teams are more able to connect the dots between what may otherwise seem to be isolated instances of manipulation. The context provided by this approach should hopefully help teams to more easily root out cross-product and cross-market manipulation in the future.
As we found in our 2025 Annual Trends in Market Abuse and Trade Surveillance report, fines associated with enforcement action reduced fairly significantly in 2025. However, this should not be interpreted as a reduction in regulatory scrutiny – rather, many regulators are in the process of implementing a more focused approach intended to prioritize the most harmful instances of abusive trading.
In Friday’s “Meet the CFTC Directors” session, Director of the Division of Enforcement David Miller echoed this sentiment. The panel was clear that this should not be interpreted as a long‑term retreat from enforcement. Instead, it reflects a reset period before a more targeted enforcement programme takes shape. Following a transitional period marked by the appointment of a new Chair and new Enforcement Direct, Miller emphasized the regulator’s intention to pursue five enforcement priorities:
Broadly speaking, regulated firms can expect to see an increase in enforcement actions against behaviors which undermine market integrity, with a decrease in “technical” cases which do not involve customer harm.
Overall, the recurring theme seen across the event was change. New technologies, new products, and new markets are all rapidly changing the face of regulatory compliance, and it is the job of compliance professionals to understand and adapt to these changes.
The challenge is no longer whether firms should adopt new technologies, but rather how to do so responsibly. AI is firmly embedded in the future of trade surveillance, but its value lies in augmentation rather than automation. In the current state of financial markets, human judgement, clear judgement, and robust controls remain essential, particularly as regulators pursue a more focused and tighter enforcement approach.
