The employment of generative artificial intelligence across different industries and sectors is old news. From healthcare, finance, and transportation to retail, cybersecurity, and education, GenAI has emerged as a fundamental tool that can be leveraged by industries to achieve everything from improved consumer experiences to more efficient in-house operations.
A little explored contribution of this technology, however, is that to material science. Although experts have highlighted that artificial intelligence is still in its pre-pubescent phase, early data points to groundbreaking transformations.
Material science has been traditionally constrained by the time it takes in discovering, testing and developing materials, in addition to high costs. Companies must comply with challenging benchmarks, from developing materials that withstand extreme environments like high temperatures and difficult corrosive conditions while following the highest levels of safety and protection. What’s more, they must also handle the decommissioning, dismantling and waste processing issues.
2024 has marked what some experts have rendered a landmark moment because of the rise in the employment of GenAI within material science. As per a recent report, AI algorithms are increasingly being used to accelerate materials discovery and development processes. “One area of focus in the prediction of new materials with desirable properties… by analyzing large datasets of material properties and structures,” the report reads.
Other obvious benefits include the optimization of the material synthesis process, swifter identification of the best conditions for producing materials with specific properties, and the aiding of researchers in designing new material applications.
As an industry pioneer in material science, Praduymna Gupta, founder and CEO of Infinita Lab, offers a fresh take on this emergent technology, and what it could mean for the future of material science.
“In emerging and critical areas such as AI-enhanced materials discovery or hypersonic materials testing, challenges lie in the development of testing methods,” Gupta affirms. In countries like the United States, these insights could not come at a more valuable time, when the vulnerabilities of the American manufacturing process are expected to be revealed as trade policies change with the incoming presidential administration.
AI’s role in democratizing research and development
One of the most overlooked aspects of AI within material science is its democratization potential. Through AI, companies can use novel compute architectures and simulation methodologies to improve the efficiency of the processes. And because these tools are openly available, all those industry professionals, laboratories, or companies who wish to expand, will be able to engage with these rising technologies.
“Accelerating deep tech application research in materials and chemical sciences by democratizing access is a groundbreaking development,” Gupta states.
Now, instead of relying on time consuming and expensive experimentation, researchers can use machine learning models that predict properties before materials are even developed in laboratories. The Graph Networks for Materials Exploration (GNoME), for instance, was proposed by Google DeepMind and uncovered 2.2 million crystal structures that were theoretically stable but unexplored in experimental settings. These compounds have now been added to the list of 48,000 that traditional experimentation had identified.
Regardless, although AI is increasingly becoming more accessible for researchers and scientists, limitations in memory and access to appropriate testing equipment pose great challenges to upcoming innovations. Hence, the importance of democratizing and inter-field cooperation.
In bringing stakeholders such as testing labs, experts, and industry leaders together to spur the development of testing methods, AI’s power is further harnessed. That is, we cannot forget that it is, after all, a tool and not an end-all-be-all solution.
Such is the case of industry pioneers like Infinita Lab. Gupta emphasizes the importance of collaboration. “In some cases, the lab is available but the capabilities have gaps that could be complemented by another lab or university research. In all cases, we start with the emerging needs in the industry, and help to answer them proactively by developing capabilities in the industry.
Beyond science: everyday considerations
It would be far too simplistic to say that AI’s impact on material science urges analysis by industry professionals and scientists alone. Material science, after all, has been the conduit for technologies as important to everyday life as renewable energy processes, optical fibres, and energy storage materials.
With AI, Gupta predicts that in-house labs will become obsolete much faster, as this technology will accelerate the evolution of materials tools and processes as much as four times. As such, contract labs with testing experts serving multiple companies have a much better chance of surviving the faster pace of demand.
Looking towards 2025, and five years into the US-China trade war, American manufacturing is likely to experience setbacks in testing, inspection and certification of infrastructure because, as they currently stand, they cannot support the rapid supply chain shifts that the current trade policies demand. Like many things, it is a matter of supply and demand.
Labs in the United States will thus struggle to adapt to the influx of demand by American companies who previously sought providers of materials science abroad. Without addressing these fundamental weaknesses, which are closer than we might have expected even six months ago, there will be an inevitable backtrack in any efforts to revitalize American manufacturing, and a possible compromise of product safety and competitive advantages.
Although it might be too soon to predict for certain, it is clear that utilizing AI within material science can give these providers, and the companies they work with, a leg up in the complicated supply chain. Will the United States, then, leverage the power of emerging technologies to satisfy the inevitable demand of in-house material testing?
Disclosure: This article includes a client of an Espacio portfolio company.
