Session (CTI)

Issue: As real-world evidence takes an increasingly important role in regulatory and health technology assessment decision making, causal inference implemented via a target trial framework is becoming more widely adopted to improve the credibility and interpretability of non-randomized evidence. At the same time, rapid advances in artificial intelligence are reshaping how healthcare data are generated, curated, and analysed. These parallel developments raise a question regarding the future of evidence generation: will causal inference continue to rely primarily on explicitly designed epidemiologic frameworks, or will it also use AI-driven systems which increasingly infer causal relationships directly from large-scale healthcare data? Overview: This issue panel will debate whether current approaches to causal inference are scalable and sustainable in an era of rapidly expanding data complexity, and whether AI can meaningfully augment (and at one end of the spectrum, potentially replace) traditional causal inference paradigms. Panelists will also discuss the implications for transparency, reproducibility, and regulatory trust. The discussion will also explore whether future "decision-grade" RWE systems will require fundamentally new architectures integrating AI-enabled longitudinal data capture, automated endpoint derivation, and dynamic protocol emulation.