Key Takeaways
- Artificial Intelligence (AI) adoption grows but confidence lags: Only 32% of scientists and informaticians feel confident in leveraging scientific data for AI initiatives, highlighting foundational data challenges across biopharma.
- Data access and integration remain top hurdles: 70% of respondents report difficulty accessing data for AI, citing issues like siloed systems, manual capture, and outdated infrastructure.
- Standardization is critical to unlocking AI’s value: With just 39% of organizations using standardized data formats and ontologies, the industry must prioritize data harmonization to move from the “Age of Data Management” to the “Age of AI.”
A new global survey from Zifo Technologies reveals that while many scientists and informaticians are beginning to integrate artificial intelligence (AI) and machine learning across the R&D, clinical, and manufacturing value chain, significant challenges remain around data infrastructure. Despite growing momentum, only 32% of respondents expressed confidence in their company’s ability to effectively leverage scientific data for AI initiatives.
The Data Readiness Survey, which included participants from more than 30 science-driven organizations, highlights both the accelerating adoption of AI and the persistent data-related hurdles facing biopharma. Among the most pressing issues: 70% of respondents reported difficulty accessing the data needed to support AI projects, pointing to ongoing problems with standardization, siloed systems, and data integration. Manual data capture, lack of interoperability, and aging infrastructure further undermine progress. Notably, most data management tools, such as electronic lab notebooks (ELNs), are ill-equipped to handle the vast, unstructured data generated in high-performance computing environments—leaving a critical gap in the AI pipeline.1,2
Are Biopharma Companies Ready to Harness the Full Power of AI?
“The current state of AI and machine learning adoption and integration within scientific organizations is still in its early stages,” said Paul Denny-Gouldson, chief scientific officer, Zifo Technologies, in the report. “A key focus is on understanding and educating scientists about the necessity of high-quality data that is well documented with metadata. This ensures the data can be effectively utilized for algorithm development and machine learning initiatives. While AI can assist in data cleaning, the primary challenge lies in establishing a foundation of clean, well described data. The speed at which this landscape changes over time will be a key indicator of progress. If minimal improvement in data accessibility for AI is observed in the near future, it might suggest that current strategies are insufficient or that this is a more protracted issue.”
