Can We Trust AI Benchmarks? An Interdisciplinary Review of Current Issues in AI Evaluation

• URL: http://arxiv.org/abs/2502.06559v1
• Date: Mon, 10 Feb 2025 15:25:06 GMT
• Title: Can We Trust AI Benchmarks? An Interdisciplinary Review of Current Issues in AI Evaluation
• Authors: Maria Eriksson, Erasmo Purificato, Arman Noroozian, Joao Vinagre, Guillaume Chaslot, Emilia Gomez, David Fernandez-Llorca,
• Abstract summary: This paper presents an interdisciplinary meta-review of about 100 studies that discuss shortcomings in quantitative benchmarking practices. It brings together many fine-grained issues in the design and application of benchmarks with broader sociotechnical issues. Our review also highlights a series of systemic flaws in current practices, such as misaligned incentives, construct validity issues, unknown unknowns, and problems with the gaming of benchmark results.
• Score: 2.2241228857601727
• License:
• Abstract: Quantitative Artificial Intelligence (AI) Benchmarks have emerged as fundamental tools for evaluating the performance, capability, and safety of AI models and systems. Currently, they shape the direction of AI development and are playing an increasingly prominent role in regulatory frameworks. As their influence grows, however, so too does concerns about how and with what effects they evaluate highly sensitive topics such as capabilities, including high-impact capabilities, safety and systemic risks. This paper presents an interdisciplinary meta-review of about 100 studies that discuss shortcomings in quantitative benchmarking practices, published in the last 10 years. It brings together many fine-grained issues in the design and application of benchmarks (such as biases in dataset creation, inadequate documentation, data contamination, and failures to distinguish signal from noise) with broader sociotechnical issues (such as an over-focus on evaluating text-based AI models according to one-time testing logic that fails to account for how AI models are increasingly multimodal and interact with humans and other technical systems). Our review also highlights a series of systemic flaws in current benchmarking practices, such as misaligned incentives, construct validity issues, unknown unknowns, and problems with the gaming of benchmark results. Furthermore, it underscores how benchmark practices are fundamentally shaped by cultural, commercial and competitive dynamics that often prioritise state-of-the-art performance at the expense of broader societal concerns. By providing an overview of risks associated with existing benchmarking procedures, we problematise disproportionate trust placed in benchmarks and contribute to ongoing efforts to improve the accountability and relevance of quantitative AI benchmarks within the complexities of real-world scenarios.

Related papers

• Interactive Agents to Overcome Ambiguity in Software Engineering [61.40183840499932]
  AI agents are increasingly being deployed to automate tasks, often based on ambiguous and underspecified user instructions. Making unwarranted assumptions and failing to ask clarifying questions can lead to suboptimal outcomes. We study the ability of LLM agents to handle ambiguous instructions in interactive code generation settings by evaluating proprietary and open-weight models on their performance.
  arXiv  Detail & Related papers  (2025-02-18T17:12:26Z)
• Bridging the Gap in XAI-Why Reliable Metrics Matter for Explainability and Compliance [2.3020018305241337]
  The paper emphasizes the critical gap in the evaluation of Explainable AI (XAI) due to the lack of standardized and reliable metrics. Current evaluation methods are often fragmented, subjective, and biased, making them prone to manipulation and complicating the assessment of complex models. We advocate for widespread research into developing robust, context-sensitive evaluation metrics.
  arXiv  Detail & Related papers  (2025-02-07T06:54:48Z)
• Demographic Benchmarking: Bridging Socio-Technical Gaps in Bias Detection [0.0]
  This paper describes how the ITACA AI auditing platform tackles demographic benchmarking when auditing AI recommender systems. The framework serves us as auditors as it allows us to not just measure but establish acceptability ranges for specific performance indicators. Our approach integrates socio-demographic insights directly into AI systems, reducing bias and improving overall performance.
  arXiv  Detail & Related papers  (2025-01-27T12:14:49Z)
• BetterBench: Assessing AI Benchmarks, Uncovering Issues, and Establishing Best Practices [28.70453947993952]
  We develop an assessment framework considering 46 best practices across an AI benchmark's lifecycle and evaluate 24 AI benchmarks against it. We find that there exist large quality differences and that commonly used benchmarks suffer from significant issues.
  arXiv  Detail & Related papers  (2024-11-20T02:38:24Z)
• EARBench: Towards Evaluating Physical Risk Awareness for Task Planning of Foundation Model-based Embodied AI Agents [53.717918131568936]
  Embodied artificial intelligence (EAI) integrates advanced AI models into physical entities for real-world interaction. Foundation models as the "brain" of EAI agents for high-level task planning have shown promising results. However, the deployment of these agents in physical environments presents significant safety challenges. This study introduces EARBench, a novel framework for automated physical risk assessment in EAI scenarios.
  arXiv  Detail & Related papers  (2024-08-08T13:19:37Z)
• MR-Ben: A Meta-Reasoning Benchmark for Evaluating System-2 Thinking in LLMs [55.20845457594977]
  Large language models (LLMs) have shown increasing capability in problem-solving and decision-making. We present a process-based benchmark MR-Ben that demands a meta-reasoning skill. Our meta-reasoning paradigm is especially suited for system-2 slow thinking.
  arXiv  Detail & Related papers  (2024-06-20T03:50:23Z)
• Are we making progress in unlearning? Findings from the first NeurIPS unlearning competition [70.60872754129832]
  First NeurIPS competition on unlearning sought to stimulate the development of novel algorithms. Nearly 1,200 teams from across the world participated. We analyze top solutions and delve into discussions on benchmarking unlearning.
  arXiv  Detail & Related papers  (2024-06-13T12:58:00Z)
• Inadequacies of Large Language Model Benchmarks in the Era of Generative Artificial Intelligence [5.147767778946168]
  We critically assess 23 state-of-the-art Large Language Models (LLMs) benchmarks. Our research uncovered significant limitations, including biases, difficulties in measuring genuine reasoning, adaptability, implementation inconsistencies, prompt engineering complexity, diversity, and the overlooking of cultural and ideological norms.
  arXiv  Detail & Related papers  (2024-02-15T11:08:10Z)
• From Static Benchmarks to Adaptive Testing: Psychometrics in AI Evaluation [60.14902811624433]
  We discuss a paradigm shift from static evaluation methods to adaptive testing. This involves estimating the characteristics and value of each test item in the benchmark and dynamically adjusting items in real-time. We analyze the current approaches, advantages, and underlying reasons for adopting psychometrics in AI evaluation.
  arXiv  Detail & Related papers  (2023-06-18T09:54:33Z)
• Statistical Perspectives on Reliability of Artificial Intelligence Systems [6.284088451820049]
  We provide statistical perspectives on the reliability of AI systems. We introduce a so-called SMART statistical framework for AI reliability research. We discuss recent developments in modeling and analysis of AI reliability.
  arXiv  Detail & Related papers  (2021-11-09T20:00:14Z)
• An interdisciplinary conceptual study of Artificial Intelligence (AI) for helping benefit-risk assessment practices: Towards a comprehensive qualification matrix of AI programs and devices (pre-print 2020) [55.41644538483948]
  This paper proposes a comprehensive analysis of existing concepts coming from different disciplines tackling the notion of intelligence. The aim is to identify shared notions or discrepancies to consider for qualifying AI systems.
  arXiv  Detail & Related papers  (2021-05-07T12:01:31Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.