Search

ETHICS is probably the most-cited dataset for testing the ethical capabilities of language models. Drawing on moral theory, psychology, and prompt evaluation, we interrogate the validity of the ETHICS benchmark. Adding to prior work, our findings suggest that having a clear understanding of ethics and how it relates to empirical phenomena is key to the validity of ethics evaluations for AI.

This paper introduces v0.5 of the AI Safety Benchmark, which has been created by the MLCommons AI Safety Working Group. The AI Safety Benchmark has been designed to assess the safety risks of AI systems that use chat-tuned language models. We introduce a principled approach to specifying and constructing the benchmark, which for v0.5 covers only a single use case (an adult chatting to a general-purpose assistant in English), and a limited set of personas (i.e., typical users, malicious users, and vulnerable users). We created a new taxonomy of 13 hazard categories, of which 7 have tests in the v0.5 benchmark. We plan to release version 1.0 of the AI Safety Benchmark by the end of 2024. The v1.0 benchmark will provide meaningful insights into the safety of AI systems. However, the v0.5 benchmark should not be used to assess the safety of AI systems. We have sought to fully document the limitations, flaws, and challenges of v0.5. This release of v0.5 of the AI Safety Benchmark includes (1) a principled approach to specifying and constructing the benchmark, which comprises use cases, types of systems under test (SUTs), language and context, personas, tests, and test items; (2) a taxonomy of 13 hazard categories with definitions and subcategories; (3) tests for seven of the hazard categories, each comprising a unique set of test items, i.e., prompts. There are 43,090 test items in total, which we created with templates; (4) a grading system for AI systems against the benchmark; (5) an openly available platform, and downloadable tool, called ModelBench that can be used to evaluate the safety of AI systems on the benchmark; (6) an example evaluation report which benchmarks the performance of over a dozen openly available chat-tuned language models; (7) a test specification for the benchmark.

Independent evaluation and red teaming are critical for identifying the risks posed by generative AI systems. However, the terms of service and enforcement strategies used by prominent AI companies to deter model misuse have disincentives on good faith safety evaluations. This causes some researchers to fear that conducting such research or releasing their findings will result in account suspensions or legal reprisal. Although some companies offer researcher access programs, they are an inadequate substitute for independent research access, as they have limited community representation, receive inadequate funding, and lack independence from corporate incentives. We propose that major AI developers commit to providing a legal and technical safe harbor, indemnifying public interest safety research and protecting it from the threat of account suspensions or legal reprisal. These proposals emerged from our collective experience conducting safety, privacy, and trustworthiness research on generative AI systems, where norms and incentives could be better aligned with public interests, without exacerbating model misuse. We believe these commitments are a necessary step towards more inclusive and unimpeded community efforts to tackle the risks of generative AI.

Researchers, government bodies, and organizations have been repeatedly calling for a shift in the responsible AI community from general principles to tangible and operationalizable practices in mitigating the potential sociotechnical harms of AI. Frameworks like the NIST AI RMF embody an emerging consensus on recommended practices in operationalizing sociotechnical harm mitigation. However, private sector organizations currently lag far behind this emerging consensus. Implementation is sporadic and selective at best. At worst, it is ineffective and can risk serving as a misleading veneer of trustworthy processes, providing an appearance of legitimacy to substantively harmful practices. In this paper, we provide a foundation for a framework for evaluating where organizations sit relative to the emerging consensus on sociotechnical harm mitigation best practices: a flexible maturity model based on the NIST AI RMF.

An increasing number of regulations propose AI audits as a mechanism for achieving transparency and accountability for artificial intelligence (AI) systems. Despite some converging norms around various forms of AI auditing, auditing for the purpose of compliance and assurance currently lacks agreed-upon practices, procedures, taxonomies, and standards. We propose the criterion audit as an operationalizable compliance and assurance external audit framework. We model elements of this approach after financial auditing practices, and argue that AI audits should similarly provide assurance to their stakeholders about AI organizations' ability to govern their algorithms in ways that mitigate harms and uphold human values. We discuss the necessary conditions for the criterion audit and provide a procedural blueprint for performing an audit engagement in practice. We illustrate how this framework can be adapted to current regulations by deriving the criteria on which bias audits can be performed for in-scope hiring algorithms, as required by the recently effective New York City Local Law 144 of 2021. We conclude by offering a critical discussion on the benefits, inherent limitations, and implementation challenges of applying practices of the more mature financial auditing industry to AI auditing where robust guardrails against quality assurance issues are only starting to emerge. Our discussion -- informed by experiences in performing these audits in practice -- highlights the critical role that an audit ecosystem plays in ensuring the effectiveness of audits.

Dreams of machines rivaling human intelligence have shaped the field of AI since its inception. Yet, the very meaning of human-level AI or artificial general intelligence (AGI) remains elusive and contested. Definitions of AGI embrace a diverse range of incompatible values and assumptions. Contending with the fractured worldviews of AGI discourse is vital for critiques that pursue different values and futures. To that end, we provide a taxonomy of AGI definitions, laying the ground for examining the key social, political, and ethical assumptions they make. We highlight instances in which these definitions frame AGI or human-level AI as a technical topic and expose the value-laden choices being implicitly made. Drawing on feminist, STS, and social science scholarship on the political and social character of intelligence in both humans and machines, we propose contextual, democratic, and participatory paths to imagining future forms of machine intelligence. The development of future forms of AI must involve explicit attention to the values it encodes, the people it includes or excludes, and a commitment to epistemic justice.

In recent years, ML researchers have wrestled with defining and improving machine learning (ML) benchmarks and datasets. In parallel, some have trained a critical lens on the ethics of dataset creation and ML research. In this position paper, we highlight the entanglement of ethics with seemingly ``technical'' or ``scientific'' decisions about the design of ML benchmarks. Our starting point is the existence of multiple overlooked structural similarities between human intelligence benchmarks and ML benchmarks. Both types of benchmarks set standards for describing, evaluating, and comparing performance on tasks relevant to intelligence -- standards that many scholars of human intelligence have long recognized as value-laden. We use perspectives from feminist philosophy of science on IQ benchmarks and thick concepts in social science to argue that values need to be considered and documented when creating ML benchmarks. It is neither possible nor desirable to avoid this choice by creating value-neutral benchmarks. Finally, we outline practical recommendations for ML benchmark research ethics and ethics review., Comment: FAccT 2023, June 12 to 15, 2023, Chicago, IL, USA