Vedran Sekara

Vedran Sekara is an Assistant Professor at the IT University of Copenhagen, affiliated with the NERDS research group. His work critically examines limitations in AI systems, focusing on data representativeness, algorithmic bias, and the predictive boundaries of machine learning in social contexts.

His research spans computational social science and network science, with emphasis on scenarios involving uncertainty and hidden variables. He demonstrates that simple models often match or outperform complex AI systems—as shown in his World Cup 2022 prediction analysis where a FIFA-ranking benchmark model achieved 81.25% accuracy in knockout stages versus AI's 75%, correctly forecasting Argentina's victory.

Sekara actively recruits PhD students for projects on ML predictability limits and algorithmic fairness, collaborating with researchers like Roby Sinatra. He contributes to the Pioneer Centre for AI and advocates for transparency in AI development, challenging commercial hype around algorithmic capabilities through public discourse on platforms like Mastodon.