Consolidating context

Tue 24 February 2026, Philippe Ropenga Artificial intelligence Estate management Graphs

You know that I have recently started working on graphs and you are as interested in legal theory as I am. Hence, it is worth taking a step back from current law to reexamine Webb v Webb from the perspective of conceptual connections. Taking this step back will enable us both to experience a practical issue that may be overlooked by people who deal with artificial intelligence and legal data processing. Let’s take a look at the following matters, starting by Webb v Webb:

• A father used funds that were on a French bank account to purchase a flat in France in the name of his son. A dispute arose between them. The father claimed that the son held the property on trust for him while the latter argued that his flat was a gift. A Court had to decide the case, but which one: the one that had jurisdiction over the dispute related to the piece of immovable property and the one that had jurisdiction over the duties of a trustee, the son, to a beneficiary, his father? It is obvious but worth reminding that before answering a question one has to know what question to answer. In one hypothesis, the father and the son are connected to the piece of property while in the other, they are connected to a fiduciary duty.¹ You may first think that these interesting questions arise only because French domestic law ignores trusts.

• You then remember a post on this blog about the connection between forced heirship and legal title without any trust.² You know that this connection is tight while the one between forced heirship and powers to manage property is loose.

• You also know that a Big Tech may not need the legal title to its users’ data to collect and exploit them for its benefit because its infrastructure enables it to possess its users’ data.³ Can you sense the connection between legal title and power, or possession without having to process the data behind the three matters above? You are cleverer than any generative artificial intelligence. These connections were not very clear to me before I started working on legal graphs.

In brief, graphs help to see and to process connections quickly while the LLM supporting a generative artificial intelligence produces outputs based on correlations that do not help the legal practitioner to understand any concrete matter.

Go Top