Are Crowdsourced AI Benchmarks Trustworthy? Experts Weigh In

The increasing reliance of AI labs like OpenAI, Google, and Meta on public platforms for AI model evaluation raises a question: are crowdsourced benchmarks truly trustworthy? While offering a broad real-world perspective, critics argue that the methodology behind these scores is fundamentally flawed. This article explores the ethical concerns and potential for manipulation in crowdsourced AI evaluation by examining expert perspectives from academics, industry professionals, and platform owners.