Nova Blog - "This is the best moment in history to become a lawyer" Carlo Rossi on AI, data and the law

Carlo Rossi: from law codes on the shelf to legal tech entrepreneur

Carlo Rossi grew up surrounded by legal culture — his mother was a lawyer, his grandfather a notary, his great-grandfather a magistrate. Old statute books lined the shelves. But what attracted him to law wasn't the adversarial courtroom; it was what his grandfather practiced: the notarial tradition of structuring business projects, designing the legal scaffolding within which entrepreneurs could build.

"What I loved most was not litigation, but — as my grandfather practised in the notarial world — the possibility of assisting in the construction and development of entrepreneurial projects." That orientation toward enabling rather than blocking would define his entire career.

He studied law and economics at Bocconi — a programme that was itself a novelty at the time, combining two disciplines that traditionally existed in separate silos. Then a large international law firm, M&A and corporate work. Then a doctorate in civil law under his Bocconi mentor. Then an LLM at New York University and a double degree at Singapore's National University, at a time when Asia was "emerging but not yet exploded." Each step was chosen to add a different layer — academic depth, international perspective, sectoral breadth.

"The professional looks out the window and sees people, trees and cars. The jurist sees movable property, immovable property, natural persons and legal persons. That framework lets you build anything."

— Carlo Rossi

What is legal tech and how is it changing the legal profession?

The conventional image of legal tech is software that automates document review or contract generation. Carlo's version is broader and more strategic: legal tech is the discipline of using technology to redesign the relationship between a lawyer and a client — from transactional document delivery to continuous, embedded partnership.

His first legal tech encounter came at 29, when two founders invited him into what became Italy's first Legal Tech company, UBDA. "It opened my eyes to the possibility of being a legal entrepreneur — something that wasn't on the dropdown menu." The hardest part, he says, was convincing his mother it wasn't a waste of his training.

The shift from old legal model to new legal model

📄

Old model: law as a product

The client receives a contract, an opinion, a deed — a static document delivered as a PDF. The relationship is transactional and episodic. The lawyer bills by the hour for producing the document.

🤝

New model: law as a continuous service

The lawyer is a legal partner who works alongside the client in real time — co-editing documents on shared platforms, providing strategic guidance before decisions are made, not after. The value is advice and foresight, not paper.

🧭

The key reframe: law as compass, not obstacle

Most founders see legal advice as "the nuisance of the day after" — something you deal with once the product is already built. Carlo's practice is to make it "the compass of the day before" — designing the business model legally from the start so that it can do things other models can't.

Is AI replacing lawyers? Why the answer is no — and why young lawyers benefit most

Carlo's answer to the question every law student is asking is direct and counterintuitive: this is the best moment in the history of the legal profession to be young and entering the field. His reasoning has two parts.

First, AI is making legal knowledge more accessible to clients — not replacing legal advice. When a client arrives having already used ChatGPT to draft their privacy policy, what changes is not that they no longer need a lawyer; what changes is that they arrive with more awareness of their rights and obligations. That awareness drives more contracts, more business relationships, more legal activity — not less. "Greater knowledge of one's own rights and obligations will help the development of more relationships, more contracts, more agreements. There is more trust."

Second, there is a structural generational inversion happening in the legal profession. Senior lawyers have built their authority on decades of accumulated knowledge and experience. AI is eroding the value of accumulated knowledge — while amplifying the value of understanding technology, digital systems, and the sectors that technology is reshaping. On those topics, a 25-year-old and a 60-year-old start from the same point.

"Trust from a client — when the topic is AI or technology — goes much more to a young person than to an older one. I wouldn't want to talk to someone aged 70 about artificial intelligence. It would worry me."

— Carlo Rossi

The practical implication for law students: specialise in areas where no one has 10+ years of experience yet. The EU AI Act, data governance frameworks, synthetic data, algorithmic accountability — these are fields where a young lawyer entering now has the same starting line as a senior partner. The race is equal and the young have the acceleration advantage.

Does EU regulation help or hurt innovation? A reframe

The standard entrepreneur's complaint about European regulation — GDPR costs time and money, the AI Act adds compliance overhead, rules slow everything down — is one Carlo hears constantly. His response doesn't dismiss the frustration but reframes the analysis.

His deeper point is about data sovereignty. The dominant AI platforms — US-based, VC-funded — are offering products below cost to accumulate training data at scale. "The competition is in being able to attract others' data without giving away your own." Every document you upload to DocuSign, every business question you ask ChatGPT, every contract reviewed in a US-hosted tool: the data accumulates in servers outside your legal jurisdiction. "In five years, what will that mean?"

The EU's response — which Carlo supports despite its complexity — is to build new legal infrastructure: the European Health Data Space, data intermediaries, data spaces for specific industries. "The concept of 'bank' is Italian — it was invented in Florence in the 15th century. We can invent new things. We certainly have the creativity."

What is synthetic data — and why it matters for AI, privacy and business models

Synthetic data is one of the most consequential recent developments in the intersection of AI and law — and one of the least well understood outside specialist circles. Carlo's firm was working on it five years ago, introduced by a startup that forced the question: what exactly are these, where can they be used, and who buys them?

Synthetic data explained

What it is

An algorithm is trained on a real dataset. It then generates a new, artificial dataset that statistically mirrors the original — same distributions, same correlations — but contains no real individual records. Each data point is invented; the aggregate faithfully reproduces the statistical reality.

Why it matters legally

Because the records are synthetic, they cannot be traced back to a real person. This makes synthetic data compliant with GDPR and other personal data regulations by design — even when the original dataset contained highly sensitive personal information. Italy's AI law and the EU AI Act now formally recognise synthetic data as a valid form of anonymisation for AI training.

Real-world example: Dutch Cancer Institute

The institute holds one of the world's largest oncology databases. Using synthetic data generation, it can share statistically equivalent datasets with external developers — who use them to build cancer-related applications and analytics tools — without exposing any real patient records. Once a solution is ready, it can be tested against the real data for validation.

Where the business opportunity is

Healthcare, financial services, smart cities, autonomous systems — any sector where data is valuable but direct sharing is legally blocked. Synthetic data creates a legal path to sharing the informational value of sensitive data without sharing the data itself. This unlocks entirely new data product categories and business models.

Why every company is a data company — and what that means for your business model

One of the most practically useful frameworks Carlo offers is deceptively simple: whatever your business makes or sells, you should be thinking of yourself as a data company. Not because data is fashionable, but because data is where long-term competitive value now accumulates.

The old business model — I give you a thing, you give me money — is what jurists call a synallagmatic contract. It was built for a world of information asymmetry: I know what it's worth, you don't, so I can charge more than it cost me. That asymmetry is rapidly disappearing. The more interesting model now is the associative contract: I share something with you, you share something with me, and value is created in the exchange itself. This is the platform model. It is also, increasingly, the regulatory model the EU is building.

"Why do we give our photos to Facebook? Because we want to share them with friends. Why do we give our data to LinkedIn? Why to Amazon? Because we get an efficient, low-cost service in return. This question gets asked too rarely. People always think in terms of a simple exchange — but the data they're providing is the real asset."

His advice to every client, regardless of sector: be the point of data reception, not just data generation. Design the reason why others would want to share their data with you. Build that into the product and legal structure from day one — not as an afterthought once the regulators come knocking.

What does Carlo Rossi look for when hiring lawyers?

Cerclex is a firm of roughly 20 people, many of whom are actively pursuing doctorates — but not in law. Carlo deliberately recruits PhD candidates in economics and computer science alongside legal specialists, because the ability to speak the language of a client's domain is more valuable than an additional year of legal specialisation.

What he screens for in interviews is harder to codify. "What fascinates me most is the person's ability to interpret their existence and their profession in an original way — to have strong opinions on certain topics that reveal free thinking." The exercise he uses: present an unresolved current legal or business problem and watch how the candidate approaches it, without any hint that there is a right answer.

The signal he finds most telling — and which he mentions with some amusement — is when a candidate says, almost incidentally, "I don't actually want to be a lawyer." For Carlo, that's not disqualifying. It's epiphanic. "I believe law can be practised in other ways, and there is space for people like that."

Frequently asked questions about AI, legal tech and data law

Is AI replacing lawyers?

No — according to Carlo Rossi, AI is making legal knowledge more accessible to clients, which increases rather than reduces demand for legal services. The risk is not replacement but irrelevance for lawyers who cannot speak the language of technology clients. Young lawyers who understand AI and data hold a structural advantage over more senior practitioners who do not.

What is synthetic data and why does it matter for business?

Synthetic data is artificially generated data produced by an AI model trained on real data. The synthetic output statistically mirrors the original but contains no real individual records — making it GDPR-compliant by design. It enables AI training in privacy-sensitive sectors (healthcare, finance) and is now formally recognised by Italy's AI law and the EU AI Act as a valid anonymisation method for algorithm training.

What is the EU AI Act and what does it mean for startups?

The EU AI Act is the European Union's regulatory framework classifying AI systems by risk and imposing compliance obligations accordingly. For startups it creates compliance costs but also market clarity and new opportunities — including explicit recognition of synthetic data as a training method, and new data governance structures (data intermediaries, European data spaces) that create entirely new business categories.

Why should every company think of itself as a data company?

Because data is where long-term competitive value now accumulates — in any sector, not just tech. Every business collects data from its operations and relationships. The strategic question is not whether to collect it but how to design the value exchange that makes others want to share their data with you. Companies that build this into their model early hold a compounding advantage over those that don't.

What skills should law students develop to stay relevant?

Specialise in new areas where no one has 10+ years of experience (AI regulation, data governance, synthetic data law). Pursue interdisciplinary knowledge — law plus economics or computer science. Learn the technical and business language of the sector you want to serve. On emerging legal topics, a 25-year-old and a 60-year-old start from the same line — and the young have the advantage in acceleration and digital fluency.

⚡ Lightning round

A failure you're grateful for?

Realising I couldn't fit the standard model of the legal profession as it existed when I graduated. I'm glad I added a new option to the dropdown menu, because none of the existing ones satisfied me.

Best advice you've received?

"In life, those who win are those who make the fewest mistakes." I interpreted it not as 'don't try things' but as 'don't persevere in approaches you've recognised as wrong.' Don't persist in a model that isn't yours.

Worst advice repeated too often?

Hyper-specialisation. Doing only one narrow thing, without maintaining curiosity about adjacent fields. It impoverishes your usefulness — and it's very boring.

Book that changed your perspective?

Blink by Malcolm Gladwell — it confirmed my belief in intuition as an essential, undervalued form of intelligence. Professionals are trained to reason analytically, step by step. But intuition can be disruptive in the best possible way, especially from people who are young and see things freshly.

One lesson for anyone listening?

Interpret your existence in an original way. Everyone has the possibility to do it. Not everyone has the courage to follow through.

🔑 Key takeaways

AI is not replacing lawyers — it's redistributing advantage within the profession. Clients with more legal awareness need more legal services, not fewer. The lawyers who lose are those who cannot speak the language of technology clients. Young lawyers who can hold the structural edge.

Specialise where no one has 10+ years of experience. On new legal topics — AI Act, data governance, synthetic data — a 25-year-old and a 60-year-old start from the same line. That equality of starting point is a massive opportunity for junior professionals willing to go deep on what's new.

EU regulation is not just a cost — it's also infrastructure. GDPR, the AI Act, data intermediaries: each creates new legal categories that enable new business models. The US builds platforms; Europe builds the rules that govern platforms. Both create opportunity.

Every company is a data company, whether it knows it or not. The strategic question isn't whether to collect data but how to design the value exchange that makes others want to share their data with you. Build this into your business model and legal structure from day one.

Uploading sensitive data to US-hosted tools is a data sovereignty decision. Every contract reviewed in DocuSign, every business question asked in ChatGPT, contributes to a data accumulation in servers outside your legal jurisdiction. Evaluate this deliberately, not by default.

Synthetic data unlocks privacy-compliant AI training. In healthcare, finance and beyond, synthetic data enables model training on sensitive domains without exposing real individual records. Now formally recognised by EU law — this is a commercial opportunity, not just a compliance workaround.

Carlo Rossi

Partner & Co-Founder · Cerclex · Data Valley

Legal entrepreneur, data law expert and Bocconi University lecturer in private law (approximately 10 years). Co-founder of Cerclex, a hybrid legal-technology firm of ~20 people that includes PhD candidates in economics and computer science alongside legal specialists. Co-founder of Data Valley, a spin-off focused on data strategy and governance for technology companies, producing industry vertical white papers and participating in EU-funded projects including the development of European foundation AI models. Holds an LLM from New York University and a postgraduate degree from the National University of Singapore. Recognised as one of Italy's leading voices at the intersection of law, AI and data.

Nova: the network built for people like Carlo

25,000+ high-potential professionals across 80+ countries — lawyers, technologists, entrepreneurs and operators building at the frontier of their fields. Membership is by nomination.

"This is the best moment in history to become a lawyer" Carlo Rossi on AI, data and the law