Augmented chemistry for better or for worse

It was thanks to a partnership of deductive thinking and experimental practice that chemistry emerged as a modern science in the 17th century.

Known as the “central science”, it informs and influences all other sciences because it is fundamental to explaining life (a timely reminder that photosynthesis is primarily a chemical process).

Chemistry is everywhere and plays a discreet role in our daily lives, satisfying everyone's basic needs. It is found in food preservation, clothing fibres, active ingredients in medicines, heating processes, and more.

As the cradle of an infinite quest for knowledge, it has given rise to process engineering, an essential technological field in the development of synthetic components that has opened up alternatives to natural chemical components.

While the contribution of chemistry and its technologies to the well-being of humanity is undeniable, the phobia that has developed around it has never been as intense as it is today.

In a very simplistic amalgamation, chemistry, its technologies and its industries are regularly stereotyped. Pollution, climate change, damage to biodiversity... Chemistry is caricatured as criminal, causing accidents and environmental and human disasters, and is often blamed for disrupting the natural order.

Added to these accusations is constant criticism of technologies derived from chemistry, particularly synthetic components, which are contrasted without nuance with the purity of natural products. It is as if chemistry were a fraud on nature, when nature itself is essentially based on chemical processes... This scientific aberration has become the dominant narrative, forgetting that it is thanks to this science that we understand the formation of natural compounds and are able to distinguish their benefits from their toxicity.

Since the Enlightenment, chemistry has been part of a movement to emancipate humanity by seeking to relegate beliefs behind knowledge. Is this enough to allay fears? NO, because as Bernadette Bensaude-Vincent observes, the science of chemistry has not been content to explain nature, it has sought to reproduce and copy the processes of matter formation¹ . And therein lies the danger.

The era of augmented chemistry

Although the opportunity to combine AI technology with chemistry dates back to the 1980s, interest was rekindled a few years ago, shortly before 2020, with a more overt introduction of AI into chemical research and applied processes². One million new molecules have been created since then³. The combination of chemistry and AI has increased the capacity to explore the molecular universe, generating a wealth of new knowledge that has been verified by experimentation.

Until recently impossible, the introduction of AI into production lines now makes it possible to control a chemical reaction from start to finish, with several beneficial consequences: optimisation of conversion rates, elimination of non-compliant synthetic products, and savings in raw materials and energy. Two years ago, Alysophil used an AI-driven API chemical synthesis control system in a real-world environment for the first time. A continuous flow chemical synthesis process was thus able to be entirely controlled by a reinforcement learning AI agent.

In addition to these positive effects, which can be described as gains, research into AI applications in molecular science on quantum computers raises hopes that its use could be extended to the field of polymers, which require greater processing capacity and are based on a larger amount of data.

Such technological innovations increase operational research capabilities tenfold and help optimise development time and overall performance. Time to market is also reduced.

Combining chemistry with AI also helps to improve facility safety through real-time monitoring of potential deviations in the production chain. The causes of anomalies can be diverse, ranging from variations in the conditions of the different reaction or transformation phases (density, temperature, pressure) to variations in the characteristics of raw materials.

This technological hybridisation points to a 21^st century chemistry that is more frugal in terms of materials and energy, more sensitive to safety issues and, more generally, more attuned to human needs. The way is now open for a chemistry of nature from which bio-inspiration will emerge, calling for a rethink of synthetic research and industrial development.

The combination of these old and new technologies naturally raises legitimate concerns. The learning methods that ultimately design the algorithmic scenarios for AI applications used in the chemical industry can be marred by a variety of cognitive biases. Admittedly, the analysis of the chemical compounds produced and the certification of their compliance with expected specifications constitute protective barriers. However, these developments are based on mathematical and algorithmic sciences combined with cognitive sciences, and it is this combination that gives pause for thought, as it fundamentally transforms the learning process. These transformations give rise to uncertainties and fears of a loss of human control, despite the experimental requirements.

Another consequence of the increased use of AI in chemistry is the threat of the human race becoming disaffected with the effort and work required to acquire knowledge. The risk is that we will succumb to intellectual laziness, which will inevitably lead to excessive dependence on AI technologies.

A science newly reborn

We are on the cusp of a renaissance in chemistry, driven by AI exploration of the molecular universe, miniaturisation of facilities and reaction control using artificial intelligence. These innovations are transforming R&D processes, increasing research power tenfold and enriching chemical engineering with hybrid technologies.

We can hope for an opportunity for our country to regain sovereignty in the supply of active ingredients and, more broadly, to shape a renewed industrial economic model.

The fears and new beliefs associated with these developments are inevitable, and we know that it is easier to break an atom than to break a prejudice. “The world will not be destroyed by those who do evil, but by those who watch and do nothing.” This maxim from Albert Einstein seems more relevant than ever when it comes to chemical engineering.

1. Bernadette Bensaude-Vincent is a philosopher and a science historian.

2. Applications of artificial intelligence to organic chemistry, the Dendral project, R.K. Lindsay, B. Buchanan, author J. Lederberg, 1980

3. ALYSOPHIL White Paper – www.alysophil.com