14:30
27/03/2025 - 14:30 - ISBA C115 -
(ULB)
Will give a presentation on :
Abstract :
Though intelligent agents are supposed to improve human experience (or make it more efficient), it is hard, from a human perspective, to grasp the ethical values that are explicitly or implicitly embedded in an agent's behaviour.
This is the well-known alignment problem, which refers to the challenge of designing AI systems that align with human values, goals, and preferences.
This problem is particularly challenging because most human ethical considerations involve \emph{incommensurable} (i.e., non-measurable and/or incomparable) values and criteria. Consider, for instance, a medical agent prescribing a treatment to a cancerous patient.
How could it take into account (and/or weigh) incommensurable aspects like the value of a human life and the cost of the treatment?
Now, the alignment between human and artificial values is possible only if we define a common space where a metric can be defined and used. This paper proposes extending the conventional Anything2vec approach to ethics, which has been successful in numerous similar and hard-to-quantify domains (ranging from natural language processing to recommendation systems and graph analysis).
This talk proposes a way to map an automatic agent decision-making (or control law) strategy to a multivariate vector representation, which can be used to compare and assess the alignment with human values. The rationale is that if an automatic agent implements a decision-making strategy, this strategy is optimal with respect to some loss function. At the same time, if the human accepts adhering to the agent strategy, this implicitly means that such an agent strategy is also optimal with respect to a weighted sum of human criteria. By making such an assumption, it is possible to recover some constraints on the weights of the human criteria that the adoption of the agent strategy implies.
The Ethics2Vec method is first introduced in the case of an automatic agent performing binary decision-making. Then, a vectorisation of an automatic control law (like in the case of a self-driving car) is discussed to show how the approach can be extended to automatic control settings.