TLDR: The Pistoia Alliance, a global non-profit, has unveiled a new initiative to ensure the safe and responsible adoption of agentic artificial intelligence (AI) in the life sciences sector. The project aims to establish industry-wide standards and protocols for AI agents, addressing concerns around trust, reproducibility, and regulatory compliance, and is supported by initial funding from Genentech.
London, UK – September 4, 2025 – The Pistoia Alliance, a leading global non-profit organization dedicated to fostering collaboration in life sciences research and development, today announced the launch of a significant new initiative focused on the safe and responsible adoption of agentic artificial intelligence (AI) within the industry. This move comes as life science professionals increasingly view agentic AI as one of the most disruptive emerging technologies expected to impact the sector over the next two to three years.
Agentic AI holds immense potential to revolutionize multi-step processes in life sciences, such as accelerating target prioritization and optimizing compound development by intelligently chaining together reasoning, tool use, and execution. However, the Alliance acknowledges that granting full autonomy to AI agents can create a ‘black box’ effect, undermining crucial aspects like trust, reproducibility, and regulatory compliance. In an industry where evidence validation is paramount, there is a pressing need for auditable agent workflows, guided by subject matter experts and approved data sources, to ensure reliable and verifiable results.
To address these challenges, the Pistoia Alliance’s new project will convene experts from pharmaceutical, biotechnology, and technology companies. Their collective goal is to establish robust standards and protocols for how AI agents should operate and interact within the life sciences ecosystem. Initial seed funding for this critical endeavor has been provided by Genentech, and the Alliance is actively seeking additional project sponsors to expand its reach and impact. The initiative aligns with the Alliance’s strategic priority to ‘Harness AI to Expedite R&D’ and complements the ongoing work of its AI/ML Community of Experts.
Robert Gill, the Agentic AI program lead at the Pistoia Alliance, emphasized the collaborative nature and anticipated benefits of the initiative. “This initiative will address the common issues we all face in integrating AI developments into a cohesive ecosystem that improves output quality,” Gill stated. “It will enable members to link standalone AI applications into a dynamic network and build workflows where multiple agents can reason, plan and act together. By becoming sponsors, organizations can act as first movers—shaping the standards, gaining early access to outputs, and ensuring they are at the forefront of the next wave of AI innovation in healthcare.”
Also Read:
- Agentic AI: Microsoft and Mastercard Pioneer Autonomous Business Systems for Commerce and Business
- UK Strengthens AI Governance: Mandatory Oversight and Safety Institute at Forefront
Early deliverables for the project include developing a life sciences and healthcare-specific agent-agent communication protocol and an AI agent standard. The outputs are expected to encompass whitepapers, guidelines, reference implementations, and scientific publications to facilitate widespread adoption. A Pistoia Alliance webinar poll of 253 professionals highlighted multimodal AI – the integration of text, image, molecular, and tabular data – as the top priority for pre-competitive collaboration, further underscoring the urgency of establishing shared frameworks for safe AI adoption. Previous research from April 2025 also indicated that resistance to change (51%) and a lack of skills (17%) remain significant barriers to innovation in the industry, while 39% of respondents expressed concern over AI models producing incorrect answers.
