Speakers:

Aviv Handler
Managing Director, ETR Advisory

Niv Bodo
Senior Manager, Risk Advisory, Deloitte

Dr. Shlomit Labin
VP, Data Science, Shield

Alex de Lucena
Director of Product Strategy, Shield
Managing compliance in an industry that has been plagued with market manipulation but slow to adopting technology, commodity trading compliance teams face unique challenges. In the past year every headline and topic of conversation started with AI-this or LLM-that. But for commodity trading compliance, what can AI help legacy systems solve?
With industry experts from Deloitte, Shield, and ETR Advisory, our speakers unraveled the regulatory landscape, explored common challenges, discussed the intersection of AI and compliance, and offered practical insights and real-world examples.
Watch the webinar to explore:

Aviv Handler
Managing Director, ETR Advisory

Niv Bodo
Senior Manager, Risk Advisory, Deloitte

Dr. Shlomit Labin
VP, Data Science, Shield

Alex de Lucena
Director of Product Strategy, Shield
Shield Insiders In-Focus
Speakers
Alex: Thank you, everyone, for attending. This is Navigating the Commodity Trading Compliance Landscape: The AI Evolution. For everyone joining online, thank you — we’re really pleased to have you, and we think we’re going to have a really good discussion today. Let’s start with introductions across the panel. Aviv, Niv, and then Shlomit — tell us who you are.
Aviv: Hi, I’m Aviv Handler. I run a small consultancy called ETR Advisory. We’re based in London, and we cover the regulation of the energy and commodities markets — wholesale energy trading and commodity trading, focused on European regulation.
Niv: Hi, everyone. Niv Bodo. Very happy to be here today. I’m a Senior Manager at Deloitte, Risk Advisory, in New York, and I focus my activity on supervision, record keeping, and surveillance of traded markets and instruments.
Shlomit: Hello, everybody. I’m Shlomit Labin, VP of Data Science at Shield Financial Compliance, with many years of experience — first in academia, then developing AI-based products, and in recent years in the financial industry as well.
Alex: Great to have everyone. I’m Alex DeLucena, with Shield. I’m our Director of Product Strategy, but I’m also our in-house SME — an SME by virtue of experience working in communication surveillance over the years. Before we jump in, a quick story about how I came to understand some elements of the commodities market. A long time ago, I had a gig for a year working for a utility as an energy auditor. My job was to go to residential real estate buildings in New York and assess how they were using their energy, and I learned a lot about the demands they had on their equipment and deliveries. The job I got after that — the one that set me on the path I’m on now — was in communication surveillance. I started at the bottom as a reviewer, looking at every email flagged by lexicon and random review. The institution I worked at had a really big energy trading presence, and it was uncanny how I could take the knowledge I’d learned walking around New York, talking to superintendents about their boilers, and apply it to this market. Intuitively, people understand that in the summer we need more electricity for air conditioning, and our utilities follow those patterns. A lot of what we trade is oriented around these very tangible use cases. But what’s less obvious to the consumer is just how complicated each of those markets is — it’s not just energy, it’s not just oil; it’s metals and so forth.
So to frame the discussion: since we have a diverse audience online, I want to start with the commodities markets writ large — who’s regulating them — and then get into the nuts and bolts of how we monitor them, best practices, things on the horizon, and a few special use cases. For anyone online, put questions in the chat and we’ll fold them in. Aviv, let’s start with you: what is this market, and what makes it different from other markets?
Aviv: First of all, the commodities market comprises people trading commodities. It’s a very old market — commodity trading goes back longer than financial trading. Even formalized commodity trading starts in London in a coffeehouse more than three hundred and fifty years ago. So it’s quite a traditional market, and that’s important for this discussion.
Who are the players, and what sort of trading happens? People trade physical commodities, and they trade financial contracts with those commodities as underlyings. Those can be traded on organized marketplaces or bilaterally. There are physical deals — I could sell you some wheat now or in a year’s time — and financial deals with those underlyings. Banks trade in those markets, as do merchants, commodity traders, producers, suppliers, and many traders in the middle. Those are often not financially regulated firms. The banks often are, but those physical traders — who will also trade derivatives, often under an exemption from financial regulation — will not be regulated. That’s an important difference in terms of the players. The last thing I’ll say is that energy, in particular gas and power, is quite different from other traditional commodities like oil, metals, and agriculture.
In terms of how those companies fulfill requirements and carry out surveillance, traditionally energy and commodity traders look like banks and financial institutions in many ways, but in some ways they lag behind in maturity of infrastructure and technology. Compliance is a bit different, partly because they’re not regulated and partly because of the very traditional nature of the business. So surveillance lags behind. Most commodities and energy firms I know do not do what I’d call deep comms surveillance for the time being — some do, but they’re mainly focused on order and transaction surveillance.
Alex: That’s a great start. Before we go deeper — you called out energy versus the other markets. Are there other distinctions, any oddities across the metal classes or agricultural markets?
Aviv: Gas and power is quite distinct, partly because it’s generally a regional business — other than with liquefied natural gas, you can’t really ship it from one place to another, whereas other commodities are more global. We shouldn’t forget emissions, which in many jurisdictions count as a commodity — certificate trading, emissions trading, and so on. Each commodity has its own oddities and works slightly differently in each part of the world. Metals work differently, coal is shipped differently, the characteristics and players are different. It’s very varied, but the energy-versus-other-commodity split is quite distinct.
Alex: And how is the way business is conducted different from traditional finance?
Aviv: If we’re talking about the traders, in many ways they resemble what you’d see in a financial institution, but some of it is more traditional, especially in non-energy commodities. A lot of commodity trading is done bilaterally — people wandering around, going to different places. At the same time, you have on-venue trading: lots of commodity exchanges, trading platforms, and other mechanisms, physically or financially settled. The mentality is very different too, including toward compliance matters.
Alex: You mentioned mentality — zero in on that and the compliance distinctions.
Aviv: The compliance departments of commodity traders — the ones who aren’t financial institutions — will generally be smaller-scale than in financial institutions. That’s changing, but at the end of the day most of these companies aren’t regulated financial institutions. Some have a regulated entity here and there, but the majority of trading isn’t carried out from a financially regulated institution, and that impacts both what you have to do and the mentality.
Alex: How might that mentality be different? Does it make them more prone to risk, or is it just a difference?
Aviv: The risks are different. In Europe, we have a specific anti-abuse regulation called REMIT — we’re about to get REMIT II any day now — which is specifically about market abuse in gas and power, physical or financial. There’s quite a high risk there, because there have been many fines: over a hundred since the rule started about ten years ago, with the fines beginning in 2015. Because fewer rules apply elsewhere, the risk of a regulatory breach is slightly lower and occurs less often, but it’s increasing. Some firms are also concerned with reputational risk.
Alex: Great overview. Niv, over to you — let’s talk about the regulatory landscape. Who’s regulating different commodities, regionally?
Niv: When we talk about energy and commodities trading, we’re talking about a few different types of instruments. There’s the exchange-traded futures and options, mostly regulated under the financial instruments regulators. We have swaps, also largely regulated under the same regulators covering financial markets, with some separate nuances. And we have the physical and spot trades, covered by national regulatory authorities in the EU, or more specifically in the US, the FERC. There’s quite a lot of overlap, because these instruments can be interconnected — you can manipulate one market using related instruments.
In the EU, we have the financial markets regulators such as the FCA in the UK, with the main regulations based on MiFID II, which includes the relevant futures, options, swaps, and commodities in its scope. Then the EU has ACER, the agency for the cooperation of energy regulators, which coordinates with the national regulatory authorities — the Italian energy regulator, the French energy regulator, and so on. Their role is to follow the REMIT regulation for wholesale energy market integrity and enforce its requirements, with authority to hand out fines and penalties.
In the US, things work a little differently. We have the CFTC, which has in its scope all the futures, options, and swaps for commodities. We have the NFA, the CFTC’s enforcement arm, similar to FINRA on the securities side. And we have FERC, the Federal Energy Regulatory Commission, which covers the physical trading. In terms of regulations: in the EU, REMIT and MiFID II, and we’re getting REMIT II as well, so there are a lot of new requirements every few years. In the US, it’s more about the Commodity Exchange Act and the Dodd-Frank Act. All these regulators are looking at insider trading and use of material non-public or inside information, conflicts of interest, and market abuse and manipulation. They all have requirements around record keeping for business communications and the monitoring of those communications. For swap dealers, especially in the US but also under MiFID II, there’s a requirement to record and maintain records of audio calls. And there’s a very specific Dodd-Frank requirement around trade reconstruction, which has been quite challenging — a moving target for most firms.
Alex: Among those regulators, what monitoring requirements have been imposed, and who’s being monitored?
Niv: Essentially, any entity or market participant trading in these products is being monitored. The monitoring includes trade surveillance for market manipulation and insider trading, and the requirement to maintain the relevant records — all the electronic communications for swap dealers, and the audio — and monitor those for any breaches.
Alex: Have we seen an increase in monitoring obligations or best practice across these populations?
Niv: The standards are definitely increasing. Firms are held to more stringent requirements, and we see it in the number of enforcements, sometimes the size of the penalties, and an additional focus on more complex schemes — an actual risk-based approach, understanding what fraudulent patterns can be implemented rather than just checking the box on a specific requirement. And we’re definitely seeing more enforcements around record keeping and communications recently.
Alex: I want to go back to what Aviv said about it being a traditional market — people walking around, negotiating in person. So who’s not being monitored, Aviv?
Aviv: It depends what you mean by monitored, and in which market. Gas and power, in terms of pure commodities in Europe, is more monitored in many ways than the other commodities, because you have REMIT and the Market Abuse Regulation. So both the NRAs (the energy regulators) and the NCAs (the financial regulators) are monitoring those markets for abuse — not necessarily communications, but trading patterns — because there’s a heavy trade-reporting requirement, particularly under REMIT. In other commodities, in Europe and elsewhere, it’s more fragmented. On-venue activity is often monitored, but not always. And there’s a lot of what I’d call traditional trading — people phoning each other, meeting each other. That’s harder to monitor, and the definition of abuse and insider trading is more difficult to nail down.
Alex: I want to bring Shlomit into this. We’ve started to talk about monitoring, which is your bread and butter.
Shlomit:Before I answer, Alex, I have a question for the team: why do you think the energy market is less heavily regulated in terms of rules and the regulators themselves?
Aviv: If I may — in terms of gas and power, it’s just a newer market. Deregulation of gas and power is quite new, thirty or forty years old. While there are financial products traded, the players usually aren’t financially regulated; under MiFID II, most use the activity exemption and don’t have to become financially regulated. Because the industry is relatively new and they’re not banks, they’re behind in organizational maturity, and because they’re not directly supervised, they’re not so heavily watched.
Shlomit: To that point, if we look at the entire world of trading, regulations are constantly in a trend of increasing. Banks were first and were put in the spotlight from the beginning. If the energy market is newer with different players, we do foresee that regulations will catch up. This is maybe a legal estimation, but from the other side, technology is catching up. Things that couldn’t be done in the past can be done now. Once, people were meeting outside the office and talking; now all communications are electronic and can be monitored. And technology now allows us to actually find the financial risks, so I expect that to be a driver for regulators to increase oversight and put more regulations and even more fines in place.
What I wonder is what will happen first: will the commodity market adopt technology and be prepared for it, or wait for the fines to come in and only then understand that technology must be adopted? I have a story from the classical financial industry. Niv talked about record keeping of voice — this is becoming a heavier requirement. The FCA, about two years ago, fined a firm (I believe one of the major banks, if I recall correctly) just for not having all of the traders’ calls surveilled; they did a random sampling. That couldn’t have happened five years ago, because transcription technology wasn’t good enough — so that expectation couldn’t be put on the table. Firms could say, “We have random sampling, that’s good enough,” and the regulator would be satisfied. But as of two years ago, the regulator said, “No, that’s not a good enough monitoring system,” and now you pay a fine just for not having a monitoring system in place. From Shield’s perspective, we’ve seen a huge increase since then in interest in voice surveillance — the thing that used to be considered nice to have. I wonder what will happen in the energy industry, because technology has evolved dramatically. Everyone experiences the capabilities of AI now — everyone is experimenting with ChatGPT, even outside work — so even less technology-savvy people and the regulators know what can be demanded. Traditional industries that in the past said “we don’t have to, we’re even afraid of adopting these technologies” will, I think, be forced to in the upcoming years.
Alex: Let’s ask Niv and Aviv first: what do you see leading the adoption of better controls — the ease of technology, or regulators pushing firms?
Niv: There are early adopters in every industry, but beyond them, usually it’s the regulatory scrutiny and pressure that comes first, and then the improvement. Another sign of maturity — something we saw in financial markets — is when firms move to an overall risk-based approach that adopts a high-watermark standard across regions and offices, rather than checking the box on local requirements. It usually comes in parallel, driven both by success stories from peer firms that adopted technology and saw improvements and cost efficiency, and by regulatory pressure, which is still the primary trigger for significant enhancements.
Aviv: I’d add a third element: alongside fines and ease of technology, there’s legislation. If you go back ten years in the European energy sector, very few firms had even a transaction surveillance system. That changed when the Market Abuse Regulation came in in 2016, which, under most interpretations, required some self-monitoring you didn’t need to do before as an unregulated firm. It started a trend of people buying transaction surveillance systems because they had to, then extending their use because it was a good idea — and that will continue with REMIT II. Recording is still quite patchy in Europe for a non-regulated firm; in the UK version of REMIT there’s a requirement to record, but under other versions there isn’t, and where there isn’t, there’s actually a difficulty around privacy law. But the many REMIT fines for market abuse are pushing people forward, and the technology gets easier to implement. So the evolution carries on — it’s just slower, but it’s happening.
Shlomit: Another factor influencing all industries is that the world is moving toward more electronic communications. What was done face-to-face is now done by call, chat, email, or application — you never go to your bank office anymore. The effect is that you have a trace for everything, so risks can now be not only monitored but also revealed later by investigators. Five or ten years ago not everything was documented; these trends make you more exposed to future discovery of financial violations and, eventually, fines. So this serves as a driver for both increased regulation and interest from companies to put controls in place.
Alex: Shlomit, what goes into developing robust detections that can surface these risks across markets generally, and across commodities specifically?
Shlomit: Traditionally, in surveillance, looking for risk indicators in electronic communication was the first step. The technology has varied — it used to be lexicon-based, and today it’s more AI-driven. The capabilities of new-age AI let you understand better, read better, and look into wider context. In commodities, that’s even more important, because the indicators are more subtle and specific — if you’re looking at a Platts window, it’s more complex and harder to catch with pinpoint phrases. And you won’t necessarily find direct risk indicators; nobody will say “front running” in a communication. So you have to look for hints of foul play — hiding activities, secrecy, repetitive communications — all around a specific trade, and raise a flag for inspection if it’s going on for a while. Because we have more advanced technology, we can look at subtler indicators, over wider context and longer periods, and pinpoint the risk itself without needing an explicit mention of it.
Alex: When we think about commodities markets, what behaviors or risks might be allowed or not allowed there that would differ from, say, equities markets?
Aviv: One of the woolliest areas is the definition of inside information, particularly in non-energy commodities. In energy, that’s now quite extensively defined under REMIT. But in more traditional commodities, it’s sometimes questionable whether a particular piece of information is inside information — and therefore whether you’re allowed to trade on it. Even in the Market Abuse Regulation, there’s a clause about commodity derivatives noting that for something to be inside information, it has to be something eventually expected to be published. People share information via brokers in ways you’d no longer do in financial markets, and there isn’t necessarily an easy solution — it’s just normal practice sometimes, and isn’t necessarily abusive.
Alex: Assuming that free flow is fine — what would be inappropriate?
Aviv: The classic example in REMIT: a power station has tripped and is out, and nobody knows it except the insiders. You’re not allowed to trade based on that information until it’s disclosed — and it has to be disclosed in a particular way, on an inside information platform under REMIT. In terms of market manipulation, it’s sometimes difficult to know whether something is manipulation. If I’m a producer and my traders are selling a large part of my produce for a particular commodity, that activity is going to move an index — and it should, because that’s what the index captures. But there’s a risk of being accused of index or benchmark manipulation. So there are interesting gray areas where something is normal practice, but its impact might lead to improper practice. That varies per commodity and per company, so even knowing what is abusive is often tough, and detecting it when you don’t know whether it’s allowed is tougher.
Alex: How would a firm evidence that they were doing the right thing, to themselves or a regulator?
Aviv: The usual way: you keep records, you have a policy, you make sure you’re doing the right thing. Another example is the oldest sort of manipulation — market cornering, where I don’t offer a commodity to the market in order to push up the price. Whether that’s abusive depends on the influence I have. You’d monitor it by looking at a combination of your inventory levels and your traders’ behavior around them.
Alex: What about cross-market manipulation — is that a meaningful risk across commodities?
Aviv: It’s a tricky one. It could be, because some commodities are interlinked — gas and power are interlinked — but a genuine cross-market manipulation is difficult to pull off and even more difficult to detect.
Niv: We see a little more cross-product manipulation, or at least detection of it, rather than cross-market. There’s definitely a lot of focus on the correlation between futures, options, swaps, and spot trades.
Alex: Are firms actively monitoring that, or taking a more reactive, periodic-check approach?
Niv: You’re starting to see more firms putting thought into designing surveillance scenarios and controls that look at these correlations, in addition to looking at the physical trades and exchange-traded activities separately. But it’s definitely tricky and fairly complex.
Shlomit: You touched on the holy grail of surveillance and compliance — the trade reconstruction effort, which has long been very challenging. This is something that, since the last burst of generative AI technology, can actually be solved now, because identifying the elements of a trade in a communication is far easier to do today. The connection that was very much needed wasn’t made until now, mainly due to technology challenges, and I expect that to change dramatically.
Alex: I think you’re saying the ability to capture these distinctions, especially across language with the LLM models coming out, gives us a clear path we didn’t have. But Aviv called this a very traditional market, and there’s a lot of what I’d call traditional data — and by that I mean not great data — that comes from those exchanges, which can be hard to incorporate to drive a meaningful signal for trade reconstruction. Given that, and given Aviv’s point about behaviors that are wrong in equities but fine or gray in commodities — how do you draw those lines, surfacing risk here while minimizing false positives there?
Shlomit: It’s the challenge of creating, as a vendor, a model that will serve one market when there are things that are common and things that are different. For the commodity market, you definitely need to develop something unique, because behaviors are not the same. The same part of a communication might be okay in a commodity market and not okay in a traditional bank. So models and the crafted behaviors should be adapted and changed according to what’s accepted in this market. You cannot just adopt an out-of-the-box bank model and put it in your commodity surveillance — and the compliance officers themselves can’t easily adopt whatever is available in the market either.
Alex: I’ll go through a few recent use cases. The first: a fine Morgan Stanley got from Ofgem. This was straightforward — wholesale energy traders on privately owned phones using WhatsApp to discuss energy market transactions. Aviv, can you talk about that case?
Aviv: This fine surprised the European energy market, because it’s a rule that was forgotten about. There isn’t a great requirement to record for non-financial European firms; the only one is in the UK version of REMIT. This isn’t a Brexit thing — when the UK transposed the regulation, they added a requirement to record communications related to activity in UK wholesale energy products, or WEPs. Many energy firms record their voice lines anyway, but not under a regulatory banner. The market had forgotten about that requirement, and suddenly this fine came out. It was interesting because it’s new territory for Ofgem — the likes of the FCA were the ones fining for this traditionally — and because the penalty was relatively high, a few million pounds, which is considered high in the REMIT stakes. Part of the reasoning was that, as an investment bank, Morgan Stanley would be expected to have that infrastructure in place to make sure off-channel communication doesn’t occur. And with REMIT II coming in, with a requirement to monitor effectively for wholesale energy products across the EU, I’d expect people to keep that in mind.
Alex: Niv, do you anticipate we’ll see more obligations to record all channels where business is conducted, the way we’ve seen with the SEC, CFTC, and FCA?
Niv: This is a topic that can’t be forgiven anymore — it’ll be enforced more and more. If you read the details of the case, it wasn’t only that these business communications weren’t recorded. There was another element: Ofgem was asking for information as part of a query, and that flagged that these communications were missing. If you look at the JPMorgan Chase case that started the whole SEC and CFTC crackdown in the US, that was one thing they took very seriously — when you don’t have the communications needed for an examination or investigation, you’re impeding that investigation, which is more serious than just the record-keeping violation. I think we’re going to see more of that focus across all the relevant markets.
Alex: Shlomit, you’ve developed detections for off-channel communications. How do we surface what we cannot see — how do we alert on what’s not there?
Shlomit: If communication occurs in an undocumented environment, you can’t monitor it directly. But people are talking, and they forget they’re constantly monitored, or assume no one will look. It used to be that WhatsApp was an unauthorized channel for banks, and it was enough that if traders even said the word “WhatsApp,” the compliance officer would call them in and ask what’s going on. So they stopped using the word within a few weeks of the system being applied, and found more sophisticated ways to talk around it — organizations kept chasing different ways to spell “WhatsApp.” But with the advanced language detection we have today, we can see the hints of someone trying to take a conversation offline — “let’s take it later,” or “ping me on…” — using words that, they think, won’t be caught under the magnitude of surveillance. Technology is catching up, and we can catch that now too. This is one of the big interests in the financial industry at the moment.
Alex: As there have been bigger weather events, we’ve seen regulators like FERC conducting audits afterward to gauge whether firms had proper controls or were manipulating the market. Do you see these events and reviews triggering a broader obligation to monitor — profanity, harassment, conduct risk, gifts and entertainment — knowing regulators may come in with broad reviews?
Aviv: I generally focus on market abuse and inside information. In terms of those other areas, that’s not happening much now, partly because international and privacy laws can make it quite hard to do, particularly in certain jurisdictions in Europe and South America. People still use traditional ways to make sure those other areas are enforced.
Alex: One more — a fine dating back to March 7, 2022. An energy trader added an extra zero to the price they were offering during a transmission balance — so instead of €245 for 500 megawatts, this sort of fat-finger error. They didn’t report it, and then they traded on it, and were fined as a result. The date matters because this was right after Russia invaded Ukraine, so you’d expect the market to be volatile — the market didn’t check itself. What behaviors might we look for to police something like that?
Aviv: The interesting background is that in REMIT, a fat-finger error is now considered inside information and has to be disclosed on an inside information website, according to the guidance of many national regulatory authorities — and there was another case around that in Finland recently. Here, though, the reason for the fine isn’t only that the disclosure didn’t happen; ACM, the Dutch energy regulator in question, has issued guidance around this. It’s that not only did they not cancel the deal or make a disclosure, but they acted on it, which sent a misleading price signal to the market — a breach of REMIT Article 5. In terms of controls, you’d look for fat-finger errors and misplaced orders: in transaction surveillance, sharp increases, big price differentials, or placing an order much larger than the best bid or offer. Comms could look at it too, to see if it was accidental or not.
Alex: And from a comms perspective, that’s why firms should also emphasize language around mistakes and errors — “don’t tell people,” “hide it” — on top of their regular front-running and spoofing detection.
Alex: Guys, we’re at time. It’s been a really great discussion — a nice back-and-forth. I think we did a good job of setting the stage and diving into some meaningful use cases across the behaviors and the detections. So thank you for your time, thank you everyone online, and have a lovely day.
Aviv: Thank you.
Niv: Thank you.
Shlomit: Thank you. Bye.
Often yes. Many commodity and energy traders operate under exemptions like MiFID II’s activity exemption, but market-abuse rules still bite. In the EU, REMIT specifically governs abuse in gas and power, physical or financial, and it has generated more than 100 fines since 2015. Recording obligations vary by jurisdiction, but the risk of a breach is real and rising, especially with REMIT II arriving.
Because behaviors aren’t the same. The exact same phrase in a communication can be perfectly normal in a commodities desk and abusive in a traditional bank. An out-of-the-box bank model dropped into commodity surveillance will misfire. Detections and crafted behaviors have to be adapted to what’s accepted in each specific market — and even the definition of abuse shifts by commodity and region.
Most energy and commodity firms today focus on order and transaction surveillance, and comms monitoring lags behind. But regulators increasingly expect it, and the recent off-channel fines show why. Communications reveal intent that trades alone can’t — whether a fat-finger price was accidental, or whether someone is steering a conversation off-channel to avoid oversight.
Yes. Morgan Stanley was fined several million pounds by Ofgem after wholesale energy traders used WhatsApp on private phones to discuss transactions. As with the JPMorgan Chase case that triggered the US SEC and CFTC crackdown, the bigger problem is that missing communications impede an investigation — which regulators treat as more serious than the record-keeping gap itself.
It’s one of the wooliest areas, especially in non-energy commodities. Under the Market Abuse Regulation, for something to be inside information it generally has to be information eventually expected to be published. In energy, REMIT defines it more clearly — the classic example is a power station tripping offline: you can’t trade on it until it’s disclosed on an inside-information platform.
It can be. Under REMIT, a fat-finger error is now treated as inside information that must be disclosed. In one 2022 case, a trader added an extra zero — €245 instead of the intended price for 500 megawatts — didn’t report it, and then traded on it, sending a misleading price signal in breach of REMIT Article 5. Controls should flag sharp price differentials and outsized orders.
Swap dealers do, under both US rules and MiFID II, and must retain audio records. For unregulated EU energy firms it’s patchier and complicated by privacy law. But expectations are rising: the FCA fined a firm roughly two years ago for relying on random sampling of trader calls — something that wasn’t enforceable before transcription technology became accurate enough to expect full surveillance.
Yes, and it’s a real shift. Older lexicon-based tools needed explicit trigger phrases, but no one writes “front running” in a chat. AI and large language models read wider context over longer periods, so they can surface subtler signals — secrecy, repetition, attempts to take a conversation offline — and make trade reconstruction far more achievable, provided the models are adapted to commodity behaviors.