Responsible AI for General-Purpose Systems: Overview, Challenges, and A Path Forward

TL;DR

This paper reviews the risks and challenges of modern general-purpose AI systems, emphasizing the need for a new approach to responsible AI that accounts for their high output flexibility and proposes a framework for future development.

Contribution

It introduces the C2V2 framework for aligning general-purpose AI with responsible AI principles and discusses how to systematically meet these requirements through system design.

Findings

Risks like hallucinations and stereotypes are more severe in general-purpose AI.

Traditional task-specific AI has lower degrees of freedom, making responsible AI easier to implement.

The C2V2 framework guides the development of responsible general-purpose AI.

Abstract

Modern general-purpose AI systems made using large language and vision models, are capable of performing a range of tasks like writing text articles, generating and debugging codes, querying databases, and translating from one language to another, which has made them quite popular across industries. However, there are risks like hallucinations, toxicity, and stereotypes in their output that make them untrustworthy. We review various risks and vulnerabilities of modern general-purpose AI along eight widely accepted responsible AI (RAI) principles (fairness, privacy, explainability, robustness, safety, truthfulness, governance, and sustainability) and compare how they are non-existent or less severe and easily mitigable in traditional task-specific counterparts. We argue that this is due to the non-deterministically high Degree of Freedom in output (DoFo) of general-purpose AI (unlike the deterministically constant or low DoFo of traditional task-specific AI systems), and there is a need to rethink our approach to RAI for general-purpose AI. Following this, we derive C2V2 (Control, Consistency, Value, Veracity) desiderata to meet the RAI requirements for future general-purpose AI systems, and discuss how recent efforts in AI alignment, retrieval-augmented generation, reasoning enhancements, etc. fare along one or more of the desiderata. We believe that the goal of developing responsible general-purpose AI can be achieved by formally modeling application- or domain-dependent RAI requirements along C2V2 dimensions, and taking a system design approach to suitably combine various techniques to meet the desiderata.