Agent-based testing: An automated approach toward artificial reactions to human behavior

Vast amounts of digital media content are consumed within algorithmically curated media environments. This renders algorithmic content curation with much societal relevance and thus highlights the urgent need for independent and resilient academic research. However, empirical investigations are sparse; not least due to unsatisfactory methodological approaches, such as case studies, observations, expert interviews, or agent-based modeling. All of these endeavors have their weaknesses, be it in terms of their validity or their reliability. Alternatively, agent-based testing has proven very suitable. Agent-based testing describes the automated conduction of a systematic variety of automated agents to experimentally vary and simulate actual human surf behavior. Technological, legal, and ethical challenges of agent-based testing approaches are discussed and practical advice is offered for an independent and resilient application.

Haim, M. (5/2019). Agent-based testing: An automated approach toward artificial reactions to human behavior. Presented at the 69th Annual Conference of the International Communication Association, Washington D.C. (content_copy)