As artificial intelligence changes health, environmental tech and robotics, experts predict longer lives, cleaner air and humanoids. Alex Zhavoronkov, PhD, a leading voice in AI-driven healthcare, recently highlighted AI's transformative possibilities during his keynote at the 2nd Annual GenAI4Pharma event at Cure.
As the CEO of Insilico Medicine, a cutting-edge generative AI drug discovery company and a Cure Collaboration Residency company, Zhavoronkov shared how AI is positioned to bring precision and efficiencies to everything from personalized medicine to environmental sustainability to human-like robots in the workforce.
His vision for an AI-enabled future within the next five years includes delaying disease onset, cleaning the air more effectively than trees and employing humanoids with superhuman capabilities in research labs. He shared with Cure details of how he thinks AI will change how we live, breathe, and work.
This interview was edited for length and clarity.
Cure: You've said that generative AI can help predict when we might develop certain diseases. Does this mean we can prevent them altogether?
Alex Zhavoronkov: Though we cannot guarantee that we won't get a certain disease, AI can be used to find measures that will help us delay or reduce the chances.
That way, if you are predicted to have a heart attack at 60, you will be more aware of heart attack issues when you're younger, so you'll monitor your heart function more frequently. You probably won't want to vigorously exercise at age 60!
Cure: How might AI help us live longer?
Zhavoronkov: Unfortunately, even if we stave off getting certain diseases, that won't increase our lifespan significantly. According to research from the University of Chicago, even if all cancers were miraculously cured, the average American's lifespan would increase by only three years.
This is why Insilico Medicine focuses on aging research, which is developing aging clocks that use a variety of data types to test gero-protective compounds.
Insilico recently released the Precious GPT series, a pioneering architecture designed to understand the biological mechanisms and the aging process for life from birth to death. It could provide high-quality biological data that mimic real-world conditions to enhance our understanding of fundamental life biology and the aging process.
Cure: You say that humanoids will soon achieve superhuman capabilities! What exactly will humanoids be able to do?
Zhavoronkov: A decade ago, the concept of a biology laboratory operating with professional precision and meticulous care — entirely without human intervention —seemed like science fiction. Today, this is a reality.
We are advancing the automation of our robotics lab, including a plan to deploy a wheeled dual-arm robot to handle primary transfer tasks, which are presently performed by automated guided vehicles (AGVs). General-purpose humanoid robots will soon be capable of performing a range of tasks.
We're exploring the use of humanoid robots to provide introductions and explanations within Insilico's AI-Powered Robotics Lab. Looking further ahead, robots with superhuman capabilities will revolutionize various fields of healthcare, such as elder care, disease diagnosis, laboratory assistance, and remote surgeries.
Cure: Insilico Medicine has an Environmental Sustainability Consortium that aims to create a material that could absorb carbon even better than trees. How does AI factor into this?
Zhavoronkov: Current technologies to capture carbon dioxide include amine scrubbing and adsorption using Metal-Organic Frameworks (MOFs) and Covalent Organic Frameworks (COFs), which are like tiny sponges with holes just the right size to capture specific types of molecules.
Scientists are also using electrochemistry, membrane technology, and biochemistry to search for new ways to capture carbon dioxide from the flue gas or directly from the air.
With the help of AI, materials with higher carbon dioxide adsorption capacity can be quickly targeted. AI can find the best building units, interaction groups, and structure topology of frameworks for carbon dioxide capture.
In the future, the application of these AI-discovered molecules for carbon dioxide capture could have a profound impact on power plants, sustainable agriculture, and oil refineries.
Cure: Back in 2016, Insilico Medicine made an early bet on Generative AI despite others' criticism and skepticism. What did you see in GenAI that your critics didn't?
Zhavoronkov: Just like with deep learning for prediction, the critics didn't understand that deep learning systems can scale and perform many tasks, even if traditional tools sometimes perform better in some benchmarks on smaller datasets.
When it came to GenAI, the critics did not understand the difference between generative reinforcement learning and more traditional approaches to synthetic data generation and drug design. To understand the real capabilities of new technology, you need to try to advance it — invest in engineering instead of trying to justify the use of older technology.