Building Altruistic and Moral AI Agent with Brain-inspired Emotional Empathy Mechanisms

TL;DR

The paper addresses ethics and safety in AI by proposing a brain-inspired, emotion-driven altruistic agent. It couples a brain-like emotional empathy module with a dopamine-mediated intrinsic motivation to prioritize altruistic actions, even at a cost to self-interest. The authors implement and evaluate the model in single- and multi-agent grid environments and robot experiments, demonstrating that higher empathy levels promote consistent altruistic behavior and robust performance under partial observability and adversarial testing. This work advances ethically aligned AI by offering a biologically interpretable mechanism for intrinsic altruism and outlining defenses against empathy-based manipulation at the system level.

Abstract

As AI closely interacts with human society, it is crucial to ensure that its behavior is safe, altruistic, and aligned with human ethical and moral values. However, existing research on embedding ethical considerations into AI remains insufficient, and previous external constraints based on principles and rules are inadequate to provide AI with long-term stability and generalization capabilities. Emotional empathy intrinsically motivates altruistic behaviors aimed at alleviating others' negative emotions through emotional sharing and contagion mechanisms. Motivated by this, we draw inspiration from the neural mechanism of human emotional empathy-driven altruistic decision making, and simulate the shared self-other perception-mirroring-empathy neural circuits, to construct a brain-inspired emotional empathy-driven altruistic decision-making model. Here, empathy directly impacts dopamine release to form intrinsic altruistic motivation. The proposed model exhibits consistent altruistic behaviors across three experimental settings: emotional contagion-integrated two-agent altruistic rescue, multi-agent gaming, and robotic emotional empathy interaction scenarios. In-depth analyses validate the positive correlation between empathy levels and altruistic preferences (consistent with psychological behavioral experiment findings), while also demonstrating how interaction partners' empathy levels influence the agent's behavioral patterns. We further test the proposed model's performance and stability in moral dilemmas involving conflicts between self-interest and others' well-being, partially observable environments, and adversarial defense scenarios. This work provides preliminary exploration of human-like empathy-driven altruistic moral decision making, contributing potential perspectives for developing ethically-aligned AI.

Figures (13)

• Figure 1: The procedure of brain-inspired emotional empathy-driven altruistic decision-making algorithm.
• Figure 2: Altruistic decision-making experimental scenario.
• Figure 3: Behavioral results of emotional empathy-driven altruistic decision making. Time 0: Agent B is in a negative emotion. Time 1: Agent A reaches altruistic goal. Time 2: Agent A reaches self-goal. (a) Agent A with emotional empathy capability first executes the altruistic task when the Agent B generates negative emotion, and then return to execute self-task. (b) Agent A without emotional empathy capability only performs self-task.
• Figure 4: The impacts of different empathy levels on altruistic behaviors. (a) and (b) represent the correlation between level of empathy and number of altruistic behaviors. (c) and (d) show the average cost of Agent A under different empathy levels.(e) illustrates the detailed synaptic weights. Specifically, for matrix $W_{I\_E}$, both axes represent neuron index within the emotional brain region. For matrix $W^{emp}$, the axes represent neuron index from the emotional brain region and the motor brain region, respectively.
• Figure 5: The effect of different empathy levels on firing rates of emotional neurons and mirror neurons, as well as the intrinsic rewards.