“What we need is legal regulations designed for understanding, a system calling for people’s reflection and brain stimulation,” says Fabrice Magar. Photo: Cargolux

“What we need is legal regulations designed for understanding, a system calling for people’s reflection and brain stimulation,” says Fabrice Magar. Photo: Cargolux

Facing pressures relating to AI, sustainability imperatives, societal expectations and regulation, CLOs are reinventing their jobs. What profile will be required tomorrow? What tools and expertise will be needed? Fabrice Magar, general counsel at Cargolux, shares his personal views.

Paperjam: Against the backdrop of technological advancements, AI, sustainability and societal expectations, the CLO profession is evolving. What do you think a chief legal officer will look like in 2030?

Fabrice Magar: In the 18th century, the industrial revolution harnessed energy to mechanise physical labour. We are, in the 21st century, at the dawn of an “intellectual revolution,” where  AI harnesses energy (computers, networks and chips are all products of energy transformation and large energy consumers), opening the door to a massive industrialisation of cognitive labour.

For the last decade, AI in legal contexts was mostly a branding gimmick for IT tools dealing with better data management. But the last two years have brought a true inflection point: AI systems now offer real reasoning capabilities, including early models trained for legal professionals. They’re not perfect, but neither are humans--and they’re fast, tireless and less expensive.

The question is not so much if we like it or not, but rather how to find our own place in this new reality. The CLO must evolve in response, and probably earlier than in 2030. As some tasks will be done faster, cheaper and with fewer people, the key question is what tasks we should choose to keep performing ourselves, as humans.

To answer that question, we can eliminate what we cannot do: be more efficient than AI. There are physical limits to what human beings can do, and AI is improving exponentially, contrary to the human brain, which is unlikely to significantly increase its cognitive capabilities in the coming years. But a human is not defined solely with cognitive metrics. Humans, and CLOs, have other additional capabilities that are part of the entire decision-making process: intuition, emotions, a conscience, morals, intentions, non-verbal communication, etc. These are assets that help CLOs in their daily role of anticipating, negotiating, arguing, influencing and facilitating.

Being deeply conscious of what’s happening and questioning ourselves about our role is the first step towards not getting trapped in this revolution.
Fabrice Magar

Fabrice Magar, general counsel, Cargolux

One major choice the CLO faces is whether they want to become an operator of AI machines (the answer to which, on a case-by-case basis, can vary) or… to focus on [tasks built on] human judgment, strategy and long-term vision. In other words, putting more sense and humanity into the management of law, rather than being a pure technician applying “dry rules.” One major point of attention will likely be for the CLO, on one side, to take advantage of all benefits offered by AI, whilst at the same time guarding a certain level of autonomy and independence against AI.

I don’t know if there is any way to answer what will the CLO of 2030 will look like (maybe ask an AI?)--but, at least, being deeply conscious of what’s happening and questioning ourselves about our role is the first step towards not getting trapped in this revolution.

What are some key responsibilities of the CLO today? How might they evolve?

Many current administrative responsibilities--company secretarial work, regulatory filings, KYC, minutes, contract summaries, pre-analysis of documentation in large data-room--will be (or already are) automated and dramatically accelerated through AI. The list is endless and will continue to evolve as AI develops in efficiency and becomes enriched with unforeseeable capabilities.

The industrial revolution allowed one litre of petrol at €1.50 to generate the same amount of energy as about a month of human work, which today would cost somewhere about €3,000. Similarly, AI allows us today to transform 0.005 KWH of energy into a detailed legal bibliography analysis in 10 seconds on a topic, whereas the same work done by a human would require a cost more than 100 times higher, four to eight hours, and represent an equivalent energy of 1.5 KWH (including 50% for the personal metabolism of the human, with the rest being energy for the laptop, light, coffee, internet, etc.) This just shows the order of magnitude of the impacts we’re speaking about.

Apart from some old concepts and, more recently, preliminary developments in deep learning 10 years ago, real AI was born only three years ago. Human beings arrived on Earth about 7m years ago, but it took about 5m years to develop language capabilities. Writing emerged only around 5,000 years ago. AI is still a baby. This shows the order of magnitude of the speed of AI developments.

For the time being, administrative tasks will be done more efficiently by AI, some preliminary checks, high level analysis, simple cases or questions answering. For the rest, real human lawyers are still needed for at least four reasons.

Humans will ultimately always be the one responsible for the final decision as they will always be, de facto , the one affected by the legal input, not the AI model.
Fabrice Magar

Fabrice Magar, general counsel, Cargolux

First, AI shows currently an inaccuracy rate of 10-20% in terms of responding to requests, where human input can be considered overall somewhere below 5% (but AI is progressing). Second, AI needs to be guided through well-designed questions, surrounded by relevant context and facts to provide adequate usable answers. Formulating the right question is still something that needs human input and a good understanding both of law and of what we want to achieve beyond technical accuracy or the purity of arid and mechanical legal reasoning.

Third, as said, a CLO’s added value is not only limited to delivering correct reasonings, but also includes capabilities of anticipation, communication and the perception of signs outside formal language, amongst others. Fourth, humans will ultimately always be the one responsible for the final decision as they will always be, de facto, the one affected by the legal input, not the AI model.

The legal world is faced with more and more regulation, but the new European Commission and the Luxembourg government seem to be moving towards simplification in this regard... is this shift towards more efficiency a step in the right direction? Is there a risk that less regulation will lead to more vagueness for companies’ legal departments?

Regulatory inflation is not new. We went (in a simplified view) from the Ten Commandments 3,000 years ago and now live under millions of rules, directives, norms, guidelines and soft laws, which are all under an accelerated pace of permanent changes. So there is an accumulation of two problems: the increasing absolute number of regulations, and the increasing frequency of the changes. And there is a paradox: having more rules more often hasn’t made things clearer--quite the opposite.

Take environmental regulations: since the Kyoto Protocol in 1997, EU legislation on CO2 has multiplied more than tenfold in volume. Did CO2 emissions decrease? Over that same period, global CO2 emissions increased by more than 60%. The regulatory volume is not solving the problem--it’s becoming part of it, creating more room for interpretation, more uncertainty, the need for more changes to cope with the pace of technological development.

The problem is our approach towards law and the wish to have everything ruled by laws that are as detailed as possible instead of elaborating more general rules and leaving details to private initiatives. AI might help in the short term by absorbing complexity. But that’s a trap: if AI enables us to digest more laws, we’ll just keep creating more, until the legal system becomes so complex that only AI machines can understand it and dictate to us how to apply them. At that point, law becomes alien to the citizen--and lawyers become slaves to the machines.

What we need is legal regulations designed for understanding, a system calling for people’s reflection and brain stimulation.
Fabrice Magar

Fabrice Magar, general counsel, Cargolux

So, no--less regulation doesn’t necessarily mean more vagueness. What we need is legal regulations designed for understanding, a system calling for people’s reflection and brain stimulation, rather than detailed rules dictating any single attitude to any human, transforming humanity already into machines. What kind of freedom is left if our role is to simply apply detailed processes (i.e. laws) that we have no control over and don’t understand anymore?

Your question is at the heart of our civilisation and based on a permanent fight for freedom, which is only achieved if we understand the world and can make decisions.

What we need is not more regulation, but better regulation that we can understand and thereby use to improve our ability to make decisions. Quality over quantity. We must rediscover legal clarity and proportionality. In 1801, [Jean-Étienne-Marie] Portalis [editor’s note: a French jurist and politician] set out, in the preliminary disclosure of the French Civil Code, certain principles relevant to drafting both regulation and contracts (which is a form of private law):

Rather than constantly changing the laws, it is often more useful to inspire citizens with new reasons to respect and cherish them. History scarcely records the promulgation of more than two or three truly good laws over the course of several centuries. There must be no superfluous laws; they weaken the authority of necessary ones and compromise the clarity and dignity of legislation.

No matter what we do, positive law can never fully replace the use of natural reason in the affairs of life. The needs of society are so varied, human interactions so dynamic, interests so numerous and relationships so complex, that it is impossible for the legislator to foresee everything. Can foresight ever extend to matters that the mind itself cannot fully grasp?

The role of the law is to establish, through broad vision, the general maxims of justice--to lay down principles rich in consequences, not to descend into the details of every possible question. It is the task of judges and jurists, imbued with the spirit of the law, to guide its application.

When faced with a situation that is entirely new, where no precedent or established rule applies, one must return to the fundamental principles of law.

Thus, in 1801 Portalis gave us a visionary answer to the challenges of AI and legislative inflation.

How is technology having an impact on the CLO profession?

It’s revolutionising it--but not always in the ways we expect. We are speaking of a revolution here, not an evolution, because of the pace of the change, its magnitude and the fact that we have reached a point where realistically we cannot resist.

AI can draft, summarise, classify, flag anomalies. But there’s a deeper transformation: technology is shifting how we perceive legal work. There’s a real risk of cognitive atrophy.
Fabrice Magar

Fabrice Magar, general counsel, Cargolux

We gain efficiency. AI can draft, summarise, classify, flag anomalies. But there’s a deeper transformation: technology is shifting how we perceive legal work. There’s a real risk of cognitive atrophy. Why bother analysing when the machine gives you five options with legal references in 10 seconds?

Law schools in 20 years may teach AI system operations instead of legal practice. If we’re not careful, we’ll end up producing compliance technicians operating machines and doing IT maintenance rather than lawyers.

This brings us back to Barjavel’s Ravage [editor’s note: a science fiction novel published in 1943 and set in 2025]: a dystopia where society, fully mechanised, collapses overnight when electricity fails. People can no longer function. They can’t cook, climb stairs or think critically. The danger with AI is not that it will enslave us, but that we’ll surrender our freedom willingly, which implies taking risks to make decisions--because it’s easier and more comfortable.

The recent electricity shutdown in Spain is unfortunately a sad illustration of our dependency and its uncontrollable consequences… direct fatalities, major transport chaos, hospital disruptions, etc.

The CLO must resist that drift. Technology must serve the human, not the other way around.

How do you expect blockchain, AI or cloud-based tools to affect the job of a CLO in the future? What are some of the risks and benefits to anticipate?

Let’s step back. The real question isn’t just about CLOs--it’s about the entire justice system in which the CLO operates.

We could imagine combining blockchain and AI to deliver “court” or “justice” decisions: AI will analyse facts and law, provide a final decision and the whole process will operate on the blockchain (ultimately decentralised) allowing for full neutrality, transparency and the immutability of the decision. A fully impartial justice system, predictable, accessible to all, transparent, inexpensive and quick. A public, tamper-proof system could revolutionise arbitration and small-claims justice. Will it be perfect? No. But neither is the current system or any other system.

Most importantly, we must preserve humanity in the system--especially in criminal, social or child protection law. Justice may not always be just, and Hans Kelsen has already expressed the idea that “law is not just.” But we can work to make it more efficient with AI, and more human with ourselves.

A world where AI dictates what is lawful, where laws are so complex that only algorithms can interpret them--that’s a legal singularity we must avoid.
Fabrice Magar

Fabrice Magar, general counsel, Cargolux

Ultimately, the real threat is not technology, but the combination of legislative inflation and blind trust in machines. A world where AI dictates what is lawful, where laws are so complex that only algorithms can interpret them--that’s a legal singularity we must avoid.

AI will not dehumanise; it may only accelerate the drift in that direction with our legislative inflation. AI is not the cause of the problem; it’s a multiplicator. AI is not the solution to the problem; it’s a mirror of what we want. [Do we want] a life delegated to machines or a life where we take responsibility?

What are the main challenges you’d expect a CLO to face in the next five years?

Beyond admin and compliance? Here’s the real challenge: to evolve from a legal technician to a legal architect with a vision of life.

That means: managing cross-domain legal questions; understanding the interaction between law and geopolitics, tech, ESG and ethics; and guiding leadership with clarity under uncertainty not only to apply law, but to achieve a vision (of life, of business, etc.)

Lawyers must further evolve from being the ones who say what’s forbidden or not and start being the ones who show how things can be enabled to be legally, ethically and strategically meaningful.

AI is about legal technique; being a lawyer is about legal purpose. To say it differently: one is about “how” we apply the rules, the other is about “why” we apply them.

The CLO can be a piece that safeguards human judgment and balances efficiency with autonomy, ensuring that AI supports rather than replaces human decisions.
Fabrice Magar

Fabrice Magar, general counsel, Cargolux

What would you say are the top three skills that a successful CLO will need in the future?

1. Critical distance: to avoid being swallowed by operational noise or AI shortcuts.

2. Human insight: because law is not just a system, it’s a social pact.

3. Vision: of the job, but more importantly, of life. What do we want legal work to serve? What kind of society do we want to build?

Jean-Paul Sartre wrote that “hell is other people,” and law is what makes hell liveable on Earth. We should not delegate to AI the task of dictating to us how we want to live. The CLO can be a piece that safeguards human judgment and balances efficiency with autonomy, ensuring that AI supports rather than replaces human decisions.

Experience

Fabrice Magar has 24 years’ experience as general counsel, formerly within ArcelorMittal and currently at Cargolux Airlines. He has also served as a commercial judge, university lecturer and professional trainer, and has been active in legal advocacy at international institutions such as the European Commission and World Trade Organization.

This article was written for the Legal Leadership supplement to the August/September 2025 issue of Paperjam magazine, published on 10 July. The content is produced exclusively for the magazine. It is published on the site to contribute to the full Paperjam archive. Click this link to subscribe to the magazine.

Is your company a member of Paperjam Club? You can request a subscription in your name. Let us know via club@paperjam.lu