Live·Open questions in longevity research
All news
AI in medicineClinical trialsTherapeutics

Cancer drug development can move faster before all-powerful AI arrives, says Ruxandra Teslo

15 November 2026· 261115001

Cancer drug development can move faster before all-powerful AI arrives, says Ruxandra Teslo

On November 13, biotech writer Ruxandra Teslo cited China's progress in cancer drug development as an example of medical advances made before the arrival of all-powerful AI. Aging researcher Kamil Pabis replied that general AI, capable of handling a broad range of intellectual tasks, could also accelerate human trials themselves.

According to a report from the IQVIA Institute, which collects data on the pharmaceutical industry, companies headquartered in China were involved in 39% of oncology trials started worldwide in 2024, compared with 5% a decade earlier. The corresponding share for US companies in 2024 was 32%. The report covers industry-funded trials from the first tests of drugs in humans through phase three, the large studies that assess benefits and safety. Trials involving companies from different countries may count toward more than one country's total.

Teslo advocates developing AI models for specific tasks in biology.

“Data and human experiments remain bottlenecks that we need to address.”

In an earlier analysis of clinical trials, she explained that early trials reveal how a drug behaves in the body and how the disease responds to it. In her proposed approach, models select molecules for testing, and early trial results provide data to train the models and select the next drug candidates. She therefore proposes running more early trials and shortening the time they take.

Large phase three trials primarily assess a treatment's benefits and safety. A drug's high likelihood of success does not, by itself, shorten the time needed to test it in humans: participants must be recruited, clinical sites coordinated, and patients followed. On November 14, Pabis proposed using general AI specifically to accelerate phase three. He argued that such AI could help with FDA approvals and trial planning, as well as handle administrative work faster and at lower cost than people.

In a coauthored article in 2024, Teslo had already proposed using AI to review requests and applications submitted to the FDA. She proposed combining automation with changes to trial regulations and practices. Teslo acknowledges the potential benefits of general AI while questioning whether it is necessary to treat cancer.

Originally published on Telegram by Ukhvat NewsView on Telegram
Sources
#cancer-drugs#task-specific-ai#phase-three-trials#fda-approvals#china-oncology