AstraZeneca's push to use artificial intelligence in designing and developing drugs has not yet produced a single clinical-stage candidate, CEO Pascal Soriot said Monday, according to Endpoints News.
That is a notably candid admission from the head of one of the world's largest drugmakers, at a moment when the pharmaceutical industry has been loudly optimistic about AI's ability to invent new medicines. "Clinical-stage" is the threshold that matters: it means a molecule has advanced far enough to be tested in people. Plenty of companies can point to AI-generated compounds on a screen; far fewer can point to one in a human trial.
But Soriot did not frame the technology as a dead end. Per Endpoints News, he suggested AI could still prove useful in boosting the success of later-stage drugs — the candidates already deep in the development pipeline, where the costs and stakes are highest. Endpoints reports that AstraZeneca is looking to AI to ease risk in late-stage trials and to speed up regulatory filings.
The distinction is worth unpacking. Discovering a new molecule is the glamorous version of AI in pharma. The less glamorous version is using it to run trials better — picking the right patients, spotting risk earlier, and assembling the mountains of paperwork regulators require faster. That second job is unsexy, but it is where drug development actually burns time and money.
Why it matters: if even AstraZeneca's AI ambitions are shifting from inventing drugs toward managing the trials for drugs it already has, the industry's most-hyped promise may be arriving in a far more modest form than advertised.