Only 24% of U.S. Pharma Executives Bet on AI for Drug Discovery, Citi Survey Finds
Citi’s latest research, released on September 7, surveyed 200 senior leaders in the U.S. pharmaceutical industry. The study found that just 24% of executives believe AI will deliver drugs that reach patients. This split between enthusiasm and caution highlights a growing disconnect in an industry where speed and cost are critical. The findings come at a time when China is rapidly expanding its AI‑driven research capabilities.
Citi Survey Reveals Skepticism Among U.S. Pharma Executives
The survey, conducted by Citi Research in early September, polled 200 executives from 15 of the largest U.S. pharmaceutical companies, including Pfizer, Johnson & Johnson, and Merck. Only 24% of respondents indicated that AI would significantly increase the likelihood of drug success. The remainder were divided: 35% were neutral, while 41% expressed skepticism or doubt. Respondents cited data quality, regulatory uncertainty, and the need for human oversight as key barriers. In contrast, China’s state‑backed initiatives, such as the 2023 AI‑Drug Discovery Program, have accelerated the pace of preclinical studies. The stark contrast between U.S. and Chinese approaches underscores a strategic gap that could reshape global drug development dynamics. Account to Citi Research.nn
Why This Gap Matters for Drug Development and Patients
The divergence in confidence has real‑world implications. First, it influences R&D budgeting; companies that trust AI may allocate more capital to algorithmic pipelines, potentially shortening discovery timelines. Second, the speed of approvals could be affected; regulators are still developing frameworks for AI‑generated compounds, creating uncertainty for both companies and patients. Third, patient access and cost are at stake; faster, more efficient development could lower drug prices, but without broad adoption, benefits may remain uneven. Finally, the global competitive landscape shifts; U.S. firms that lag in AI
Despite the hype, a 2022 study found that AI‑generated molecule libraries contain 30% more chemically unstable compounds than those designed by human chemists.
