The race is on in the world of drug discovery, and artificial intelligence (AI) is leading the charge. But as these smart machines speed up the process, they’re also revealing a bottleneck: our outdated data systems.

Paul Belcher at Cytiva sees AI transforming hit identification from empirical screening to predictive design. It’s quicker, more efficient, but not infallible. AI still needs robust data—and lots of it—to make accurate predictions and avoid bias.

The challenge is finding high-quality data that isn’t just a success story. ‘We often joke that there should be a journal of negative data,’ he says. But such data remains hard to come by, buried in lab notebooks or unused for future research.

Adding to the pressure is the risk of fabricated data, which could skew AI models. This isn’t just about ethics; it’s about building reliable tools that can save lives. As AI gets smarter, we must get our act together and ensure the data it relies on is as accurate as possible.