AI and engineering biology: three lessons for scale-up from the UK-US symposium

Yet one theme emerged consistently throughout the symposium: the future success of AI-enabled engineering biology will depend on far more than increasingly powerful algorithms. Real-world impact will require the foundations that allow AI to deliver value, including high-quality data, trusted standards and effective routes from research to to scale-up. 

This may prove to be one of the defining challenges, and opportunities, for the sector over the coming decade. 

Why data matters for AI-enabled engineering biology  

Across the discussions, participants highlighted that high-quality data is the critical enabler for AI-enabled engineering biology. Existing biological datasets often remain fragmented, difficult to access, poorly annotated or insufficiently connected. Many contain observational data but lack the intervention and perturbation datasets required to move from correlation towards prediction and causation. 

The question is no longer whether we have enough biological data. It's whether we have the right data. 

The symposium highlighted the importance of designing future research programmes with AI in mind from the outset. That means generating and recording richer metadata, improving reproducibility, establishing common standards and making it easier to connect datasets across organisations, disciplines and national borders.  

Participants also recognised that unlocking the full value of engineering biology data will require new approaches to data governance and sharing. While industry, academia and public programmes generate vast amounts of potentially valuable data, concerns around intellectual property and commercial sensitivity can limit accessibility. Developing approaches to anonymisation that protect confidential information while retaining the underlying value of datasets will be critical if AI models are to benefit from a broader and more representative evidence base. 

Participants also highlighted the potential value of negative data and the often-overlooked know-how generated through unsuccessful experiments. Failed experiments are rarely published, yet they can provide important insights into biological behaviour, process limitations and scale-up challenges. Capturing and sharing this knowledge in appropriate ways could help improve predictive models, reduce unnecessary duplication and accelerate learning across the engineering biology community.  

Why scale-up remains a challenge for engineering biology  

The symposium reinforced a familiar reality for biomanufacturing: success in the laboratory does guarantee success at commercial scale.  

Engineering biology has generated remarkable innovations in recent years, but translating promising biological designs into commercially viable products continues to be one of the sector's biggest challenges. Cells behave differently at industrial scale, manufacturing environments introduce new variables and economic realities become more significant. In many cases, we still lack sufficient understanding of why processes succeed, fail or perform differently during scale-up.  

AI could help address some of these challenges. Participants discussed opportunities for digital twins, advanced process monitoring, predictive modelling and self-optimising bioprocesses to improve process understanding, reduce development risk, accelerate scale-up and improve manufacturing efficiency. By enabling organisations to better predict outcomes and identify issues earlier, AI has the potential to move biomanufacturing from a largely empirical process towards a more predictive and data-driven discipline. 

These capabilities depend on access to high-quality process data and robust standards for collecting, sharing and interpreting it. If the engineering biology community can establish these foundations, AI could become a powerful tool for addressing one of the sector's most persistent challenges: translating innovation from the lab to reliable, commercial-scale manufacturing. 

Why UK-US collaboration matters for engineering biology  

One clear message from the symposium was the recognition of the complementary strengths held by the UK and the United States. Participants consistently described the relationship as one of mutual advantage rather than competition.  

The UK brings strengths in engineering biology research, standards development, collaborative networks, biofoundry infrastructure and unique biological datasets. The US offers world-leading AI capabilities, computational infrastructure, venture investment and experience in translating research into commercial success.  

Bringing these capabilities together could accelerate progress, but significant barriers remain. These include fragmented funding structures, intellectual property challenges and limited mechanisms for sustained bilateral collaboration. 

Building the foundations for AI-enabled engineering biology  

Perhaps the most important takeaway from the symposium is that the future of AI-enabled engineering biology won't be determined by algorithms alone. 

It will depend on generating better data, establishing trusted standards, sharing knowledge effectively and creating pathways that translate innovation from discovery to deployment. 

As engineering biology enters an increasingly AI-enabled era, the opportunity isn't simply to apply AI to biology, but to create the conditions that allow AI to deliver meaningful impact. It's about whether we can build the data, standards and manufacturing foundations needed to turn scientific breakthroughs into real-world impact at scale. 

CPI works with organisations across engineering biology and biomanufacturing to help turn innovative ideas into scalable, real-world solutions. To explore how CPI can support your innovation journey, get in touch with the team.

Let's innovate together.