Developing a drug is a complex process that often fails. It begins by identifying a target, such as a protein or gene, associated with a disease. Researchers then search for a molecule that can either block or enhance the target’s activity safely. This can involve screening as many as 1 million compounds before selecting just one or two promising candidates. Software can help to identify such molecules. But generative artificial intelligence (AI) can dream up entirely new ones to test. BCG, a consultancy, estimates that about 65 AI-inspired molecules are currently being tested on humans.