How life sciences can make AI a driver of discovery, not just efficiency

Each year, the Pistoia Alliance’s Lab of the Future survey takes the pulse of how far life sciences R&D has progressed in its digital transformation journey. The 2025 survey gathered insights from more than 200 global experts spanning pharma, biotech, academia, software and non-profits. The data captures where laboratories are investing, what challenges they face, and actionable steps to progress technology adoption.

This year’s results show a subtle but significant shift from previous surveys. Scientists now see the main benefit of digitizing the lab as accelerating innovation and new breakthroughs, whereas in 2024 and 2023 the industry was focused on improving the efficiency of R&D. This evolution shows how life sciences companies are moving beyond using digital tools to simply speed up workflows and are now focused on enabling better science. The Alliance’s findings are corroborated by other industry surveys, such as those by McKinsey and IQVIA, which also show organizations are prioritizing innovation and discovery as primary objectives for AI, rather than administrative or repetitive tasks.

While this vision of using technologies like AI to drive discovery is potentially transformative, it is also ambitious. So, how are labs laying the groundwork to realize this goal?

You can’t leap to the lab of the future without first improving the lab of the present

The survey shows renewed investment in core systems that deliver the best of both worlds: improving efficiency today while creating an environment where AI can thrive tomorrow. For example, electronic notebook use jumped to 81%, up from 66% in 2024, confirming a decisive move away from paper-based processes and free-text formats such as email or Word. Cloud platform use has climbed from 70% to 80%, driven by instrument vendors moving software online and by fewer security concerns as the benefits of scalability, accessibility and collaboration become clearer. These foundational technologies now form the backbone of modern R&D, ensuring that data can be shared, analyzed and reused – and captured in a format that makes it ready for AI applications.

At the same time, enthusiasm for some other high-profile technologies has cooled. Adoption of digital twins has fallen to 17% (from 23%), expected use of quantum computing has slipped to 18% (from 20%), and wearables to 35% (from 41%). These trends suggest a more disciplined approach to innovation. Organizations are less concerned about chasing every new technology that emerges, and are concentrating on strengthening their digital foundations first, with a focus on AI.

AI maturity brings progress and new pressures

Given AI’s potential to enhance workflows in areas such as target identification, molecule design and predictive modelling of complex biological systems, it’s no surprise AI remains the number one investment priority for the third-year running, cited by 63% of respondents.

However, as AI use cases expand, new challenges are emerging. Skills shortages have become one of the biggest barriers to adoption, with more than a third (34%) of respondents citing a lack of expertise, up from 23% in 2024. AI is evolving faster than any other lab technology, with new models such as multimodal and agentic AI emerging every few months. This pace of change makes access to skills and expertise increasingly critical if companies are to keep up with AI developments.

Data quality also remains a concern. More than a third (38%) pointed to low-quality or poorly curated datasets as a blocker to AI – although this is an improvement from 52% in 2024, suggesting that data infrastructure investments are beginning to pay off. Regulatory uncertainty has also fallen, with just 9% now citing it as an AI barrier in the wake of new laws such as the EU’s AI Act – down from 20%. Meanwhile, a quarter of respondents said AI still isn’t fully trusted, underlining the need for continued focus on validation and transparency.

How to turn promise into progress

The conversation around AI in life sciences is shifting from “can we use it to speed work up?” to “how do we use the technology to responsibly and effectively drive discovery?”. For companies looking to make that leap, the path forward lies in strengthening skills, data governance and collaboration across the scientific community.

Researchers want practical, shared resources to guide AI implementation:

• More than half (55%) call for best-practice use cases, business cases and “how-to” guides. A similar proportion want opportunities to collaborate with other organizations to share knowledge and risk.
• 45% want access to AI/ML educational courses and skills training, underscoring the desire for continual learning as AI technologies evolve.
• Ethical and regulatory guidance also remains high on the agenda. More than a third (35%) of respondents requested clearer frameworks for responsible AI use, and around a third want freely available data governance principles to ensure consistency and trust in shared datasets.

These findings reinforce that no single organization can advance AI in isolation. Shared frameworks, open educational resources and cross-industry partnerships are essential to translate digital investment into better science. Initiatives such as the Pistoia Alliance’s FAIR Toolkit, Future Labs Evolution Community, AI/ML Community of Experts, and new training programs show how collective action can close the skills gap and support responsible innovation.

With the right skills, standards, and opportunities for collaboration in place, AI can evolve from automating routine tasks to actively driving scientific discovery.