In a world where trust can fracture, where families blur at the edges, and intentions don’t always match outcomes, the notion of an “algorithmic trustee” arrives like a promise wrapped in clean code with no bias, no emotion, no sleepless nights. It sounds safe and easier, with many, I’m sure, seduced by the comfort of certainty.
However, real life rarely follows a clean line. So, in a future AI world where outcomes could be calculated with machine precision, can we confidently believe that AI can replace a human Trustee?
Consider the difference between fairness and justice. An algorithm, trained on rules, will pursue fairness relentlessly: equal inputs, equal outputs, symmetry above all else. However, the value add for a human trustee is that they understand that justice is dependent on circumstances and that equality is not always equity. Sometimes treating everyone the same can, in fact, be the most unfair thing of all.
An algorithm can optimise distribution; it can ensure compliance; and it can “probably” in the future remove the risk of human error. However, it cannot feel the weight of context. A human trustee remembers that one beneficiary was always the quiet one, the one who never asked for more, but perhaps needed it most. They remember the years that never made it into formal records: the caregiving, the estrangements healed in private, the sacrifices made without witness.
You are within your rights to counter my opinions above and offer up the letter of wishes as a solution. Of course, the LOW is a key document for any Trustee but there is so much of family life which lives in what hasn’t been written down, and of course “change” is always inevitable. Families evolve in ways no static model can predict. Relationships fracture and mend, and values and priorities shift with time and what felt like the right decision five years ago may feel impossibly wrong today.
Clients don’t choose trustees because they are flawless. Clients choose them because they know us or are willing to try, and this relationship can take years to build and preserve, often through generations. We respect the hierarchy of a family office; we genuinely care for the matriarch or patriarch that passes, we will visit your family business and sample the goods with you, we will hold your hand in finding trusted advisers, we can offer a trust fund in secure jurisdictions to protect your assets. We provide meaning, stability, honesty, and above all, trust.
This is not to say there is no place for AI as there is beauty in its precision and its ability to process vast landscapes of information without fatigue. Right now, it can support trustees, guide them, illuminate risks and inconsistencies invisible to the human eye and increase the speed and efficiency with which Trustees can make decisions but for now, it should remain just that: a guide.
Because when it comes to families, to legacy, to the fragile and deeply human act of deciding what happens next, my current stance is that logic alone is not enough to build the necessary level of trust to understand the family dynamics or replace the core fiduciary duties of a responsible Trustee to the Trust beneficiaries.