Problems Identified in “Toward Trustworthy AI Development: Mechanisms for Supporting Verifiable Claims”

Drawn from: Brundage, Miles, Shahar Avin, Jasmine Wang, Haydn Belfield, Gretchen Krueger, Gillian Hadfield, Heidy Khlaaf, et al. “Toward Trustworthy AI Development: Mechanisms for Supporting Verifiable Claims.” arXiv, April 20, 2020.

“This report suggests various steps that different stakeholders in AI development can take to make it easier to verify claims about AI development, with a focus on providing evidence about the safety, security, fairness, and privacy protection of AI systems. Implementation of such mechanisms can help make progress on the multifaceted problem of ensuring that AI development is conducted in a trustworthy fashion.”

2.1 Third-Party Auditing
The process of AI development is often opaque to those outside a given organization, and various barriers make it challenging for third parties to verify the claims being made by a developer. As a result, claims about system attributes may not be easily verified.

2.2 Red Team Exercises
It is difficult for AI developers to address the “unknown unknowns” associated with AI systems, including limitations and risks that might be exploited by malicious actors. Further, existing red teaming approaches are insufficient for addressing these concerns in the AI context.

2.3 Bias and Safety Bounties
There is too little incentive, and no formal process, for individuals unaffiliated with a particular AI developer to seek out and report problems of AI bias and safety. As a result, broad-based scrutiny of AI systems for these properties is relatively rare.

2.4 Sharing of AI Incidents
Claims about AI systems can be scrutinized more effectively if there is common knowledge of the potential risks of such systems. However, cases of desired or unexpected behavior by AI systems are infrequently shared since it is costly to do unilaterally.

3.1 Audit Trails
AI systems lack traceable logs of steps taken in problem-definition, design, development, and operation, leading to a lack of accountability for subsequent claims about those systems’ properties and impacts.

3.2 Interpretability
It’s difficult to verify claims about “black-box” AI systems that make predictions without explanations or visibility into their inner workings. This problem is compounded by a lack of consensus on what interpretability means.

3.3 Privacy-Preserving Machine Learning
A range of methods can potentially be used to verifiably safeguard the data and models involved in AI development. However, standards are lacking for evaluating new privacy-preserving ma- chine learning techniques, and the ability to implement them currently lies outside a typical AI developer’s skill set.