Presentation (CTI)

OBJECTIVES: To investigate the potential applications and limitations of generative Artificial Intelligence (AI) to optimize Health Technology Assessment (HTA) processes.
METHODS: A narrative, use-case-driven analysis was conducted based on peer-reviewed literature and grey literature, including reports from international organizations and institutions such as the FDA, NICE, and Cochrane, published between June 2020 and January 2025. Applications of generative AI currently used in market research were mapped and translated to core HTA domains, including systematic literature reviews, real-world evidence generation, and health economic modeling. Key considerations related to prompt engineering, transparency, ethics, and governance were qualitatively assessed.
RESULTS: Transparency and rigor in AI-assisted evidence analysis and decision support was consistently highlighted as a critical enabler of credibility and trust, in line with international guidance such as the NICE position statement on AI in evidence generation. Key limitations were also identified, including risks of bias, lack of interpretability, and the potential for hallucinated or misleading outputs, reinforcing the need for human oversight and clear documentation of assumptions. More specific prompts were consistently associated with more accurate and interpretable results. Although not yet fully explored within HTA, an emerging trend toward the use of generative AI agents in market research was identified in the literature. These agents show potential to support a range of HTA activities, including literature summarization, synthetic data generation, accelerated evidence-based decision-making, and scalable stakeholder engagement, such as enabling interview-level depth in public consultations.
CONCLUSIONS: Generative AI can act as a valuable decision-support tool across the HTA workflow when applied transparently. Its use should aim to enhance rather than replace human judgment, with clear governance frameworks.