Legal and ethical implications of AI-based crowd analysis: the AI Act and beyond.

The increasing global population and the consequent rise in crowded environments have amplified the risks of accidents and tragedies. This underscores the need for effective crowd management strategies, with Artificial Intelligence (AI) holding potential to complement traditional methods. While AI offers promise in analysing crowd dynamics and predicting escalations, its deployment raises significant ethical concerns, regarding privacy, bias, accuracy, and accountability. This paper investigates the legal and ethical implications of AI in automated crowd analysis, with a focus on the European perspective. We examine the effect of the GDPR and the recently accepted AI Act on the field. The study then delves into remaining concerns post-legislation and proposes recommendations for ethical deployment. Key findings highlight challenges in notifying individuals of data usage, protecting vulnerable groups, balancing privacy with safety, and mitigating biased outcomes. Recommendations advocate for non-invasive data collection methods, refraining from predicting and decision-making AI systems, contextual considerations, and individual responsibility. The recommendations offer a foundational framework for ethical AI deployment, with universal applicability to benefit citizens globally.