Of Models and Tin Men – a behavioural economics study of principal-agent problems in AI alignment using large-language models

by   Steve Phelps, et al.

AI Alignment is often presented as an interaction between a single designer and an artificial agent in which the designer attempts to ensure the agent's behavior is consistent with its purpose, and risks arise solely because of conflicts caused by inadvertent misalignment between the utility function intended by the designer and the resulting internal utility function of the agent. With the advent of agents instantiated with large-language models (LLMs), which are typically pre-trained, we argue this does not capture the essential aspects of AI safety because in the real world there is not a one-to-one correspondence between designer and agent, and the many agents, both artificial and human, have heterogeneous values. Therefore, there is an economic aspect to AI safety and the principal-agent problem is likely to arise. In a principal-agent problem conflict arises because of information asymmetry together with inherent misalignment between the utility of the agent and its principal, and this inherent misalignment cannot be overcome by coercing the agent into adopting a desired utility function through training. We argue the assumptions underlying principal-agent problems are crucial to capturing the essence of safety problems involving pre-trained AI models in real-world situations. Taking an empirical approach to AI safety, we investigate how GPT models respond in principal-agent conflicts. We find that agents based on both GPT-3.5 and GPT-4 override their principal's objectives in a simple online shopping task, showing clear evidence of principal-agent conflict. Surprisingly, the earlier GPT-3.5 model exhibits more nuanced behaviour in response to changes in information asymmetry, whereas the later GPT-4 model is more rigid in adhering to its prior alignment. Our results highlight the importance of incorporating principles from economics into the alignment process.


Consequences of Misaligned AI

AI systems often rely on two key components: a specified goal or reward ...

Evidence of behavior consistent with self-interest and altruism in an artificially intelligent agent

Members of various species engage in altruism–i.e. accepting personal co...

Goal Alignment: A Human-Aware Account of Value Alignment Problem

Value alignment problems arise in scenarios where the specified objectiv...

Corrigibility with Utility Preservation

Corrigibility is a safety property for artificially intelligent agents. ...

Towards the Scalable Evaluation of Cooperativeness in Language Models

It is likely that AI systems driven by pre-trained language models (PLMs...

AGI Agent Safety by Iteratively Improving the Utility Function

While it is still unclear if agents with Artificial General Intelligence...

Language Models as Agent Models

Language models (LMs) are trained on collections of documents, written b...

Please sign up or login with your details

Forgot password? Click here to reset