Even if without question the implementation of AI in the pharma industry is experiencing great developments, there are still big challenges limiting its potential in the R&D process. For instance, the biggest challenge is that most of the above mentioned and identified start-ups focus only on the preclinical stages of drug development. Their goal is, in fact, to identify better drug candidates and speed up the overall development while using resources more efficiently, which is naturally the area where AI can play out its strengths.
However, an acceleration of the preclinical development does not influence the number and length of clinical trials required, which today make up more than 50% of the total costs and face increasing regulatory changes and requirements. This problem is based on the fact that even promising test results in an in-vitro environment often show undesirable effects when transferred to in-vivo testing in mice and later human patients due to the sheer complexity of the interaction between drug and body.
Of course, there are also companies that are working on improving the transferability of in-vitro or in-silico results to in-vivo tests. But the regulatory approval or a potential substitution in the future for actual clinical trials has still a long way ahead.