Can AI predict the (regulatory) future?

Pharmaceutical companies are increasingly using artificial intelligence to predict how regulatory authorities will respond to drug approval applications. The technology analyzes past decisions from regulatory bodies like the FDA and identifies weaknesses in documentation before submission. According to an FDA study from 2014, half of new drug approval applications failed on first attempt, and successful resubmissions took a median of 435 days. Companies such as Peer AI and Novo Nordisk are developing AI platforms to optimize the documentation submission process and adapt requirements for different regulators. The FDA itself launched the generative AI tool Elsa in June 2025 to accelerate the evaluation of submissions. The goal is to reduce approval timelines and lower failure risk while creating integrated AI systems spanning from clinical trials to final approval.