Rapid deployment of AI in scientific pursuits is creating an assembly line of discoveries that is revolutionising the world, writes Satyen K. Bordoloi.


“The assembly line is a revolutionary manufacturing process that transformed production methods during the Industrial Revolution,” ChatGPT began. “It involves a product moving along a conveyor belt or track, with workers or machines at each station performing specific tasks to gradually assemble the final product.

“This method significantly increased efficiency, reduced costs, and enabled mass production.” As this article discusses two seemingly disparate subjects—Assembly Line (AL) and Artificial Intelligence (AI)—I thought it’d be interesting to let one – AI, summarise the other – AL.

The 19th century was an age of immense scientific progress. The 20th century became the age of applying those discoveries for the benefit of mankind. Incrementally, at times, rapidly at others, science gave birth to an era even our most imaginative minds couldn’t conjure up in science fiction. One of the most significant contributors to this 20th-century renaissance was mass production. The Assembly Line, which enabled this mass production, has made it possible for billions to live more comfortable and longer lives than past kings.

Henry Ford revolutionised car production with the assembly line, paving the way for mass production in the 20th century (Image: Wikipedia)

Like the industrial world a century and a quarter ago, the scientific world is experiencing its assembly line moment thanks to Artificial Intelligence streamlining scientific processes much like in production lines. To understand how that is happening, we must first understand the assembly line in mass production.

A World Remade by Assembly Line: It began with Henry Ford’s car production. Today, almost everything you use—phones, fridges, toys, food, clothing, furniture, medicines, soap, paper, pens, balls, cycles, bricks, pipes, etc.—has been entirely or partly produced on an assembly line.

The process is simple: scientists discover something useful, and engineers figure out how to make it in bulk via an assembly line. Imagining an assembly line for a car is easy. Hence, let’s try to understand it via something not so obvious: pharma companies’ mass production of vaccines.

DeepMind’s AlphaFold AI system predicted the structures of over 200 million known proteins

The assembly line of vaccines looks like this: First, cell cultures are prepared and infected with the virus. The mixture is then harvested, and the virus is separated from the cells. The harvested virus undergoes purification to remove impurities and is inactivated or weakened to ensure it triggers an immune response without causing illness. This treated virus is mixed with stabilisers and adjuvants to formulate the final vaccine. The entire process is streamlined on an assembly line for efficiency, with the final packaging occurring on another assembly line.

The assembly line has remade the world, with nothing untouched by its efficient production, except one: scientific discoveries. For the longest time, it seemed impossible; how can you make an assembly line for intelligence? It turns out AI is just that—transferring intelligence from biological beings to mechanical objects to make it ubiquitous. Intelligence itself has become an assembly line production, meaning that everything that follows, including inventions and discoveries, can now have its own assembly line.

Every field in the world has had its assembly-line moment. Now, science does, too.

Even medicines and vaccines are mass-produced in assembly line production

The Assembly Line of Discoveries by AI: Two years ago, DeepMind, a Google subsidiary, stunned the scientific world when their AI, AlphaFold, predicted the structures of over 200 million known proteins. This breakthrough would have taken centuries using traditional methods.

The impact was so profound that the team won the Nobel Prize in Chemistry in 2024—an unprecedented recognition given the short time since the discovery. The scientific community will take decades to utilise these findings’ potential fully.

DeepMind’s AlphaFold process mirrored an assembly line’s efficiency in predicting protein structures. It began with data acquisition and curation, akin to sourcing raw materials for vaccine production. Massive amounts of protein sequence data were gathered and processed to prepare a high-quality dataset. This dataset was then used to train the AI model, a process similar to inoculating a culture of cells to produce a vaccine.

Once trained, the model predicted the 3D structures of proteins, analogous to harvesting viral particles. These predictions were subsequently validated and refined through rigorous testing, similar to purifying the vaccine to remove impurities.

The protein folding process as predicted by AI, which would have taken centuries using traditional methods
(Image: Wikipedia)

The validated protein structures were integrated into a comprehensive database, akin to formulating a vaccine by combining purified viral particles with stabilisers and adjuvants. This database was made publicly accessible, allowing researchers worldwide to utilise this information.

The AI Assembly Line for Scientific Discoveries: Just as the assembly line streamlined manufacturing processes, AI accelerates scientific exploration by processing vast amounts of data at incredible speeds. This rapid analysis enables researchers to identify patterns, generate insights, and make breakthroughs faster than traditional methods.

AI also automates routine tasks, allowing scientists to focus on creative problem-solving and innovative thinking. AI enhances the precision of scientific experiments and analyses by reducing human error and ensuring consistent, accurate results. Its scalability allows for the analysis of complex datasets, making it possible to tackle large-scale scientific problems and conduct extensive simulations and experiments that were once impractical.

AI is transforming scientific research, bringing efficiency and precision to complex processes

AI also fosters interdisciplinary collaboration by integrating knowledge and techniques from multiple scientific disciplines. This symbiosis of ideas drives innovation and leads to discoveries. Just as the assembly line could predict production outcomes based on standardised procedures, AI’s predictive capabilities can forecast scientific outcomes, model potential scenarios, and guide researchers toward the most promising avenues for discovery.

Even the allocation of scientific resources can be optimised with AI, ensuring that time, funding, and materials are used efficiently, leading to cost-effective and sustainable research practices.

The Assembly Line Effect on Discoveries: The assembly line of science will have a similar effect on discoveries as it did on producing goods. Over a century ago, only a cobbler trained in making shoes could make them. A single human can indeed make a shoe all by herself even today. But how many of us can gather all the engineering knowledge and manufacturing skills required to build an entire car? Is there one person in the world who can make a mobile phone all by themselves, including everything that goes into it?

It is no wonder that the assembly line started at the beginning of the age of creating complex goods like cars for the masses. In the same vein, when it comes to scientific inventions and discoveries, the simplicity of the days when a Newton or Einstein discovered fundamental principles is over. Like manufacturing complex equipment, science has become a complex endeavour where every scientist discovers a tiny bit that adds to solving the big puzzle.

In such a scenario, AI becomes the perfect assembly line, not just to understand this complexity and the position of each bit that goes into it but also to create your tiny bit and send it along the assembly line for others to pick up and fit into the more significant finding.

Like assembly lines ensured one didn’t have to know how to make something in entirety, the assembly line of AI in the scientific world means that young researchers can leverage AI to accelerate scientific discoveries and innovation without knowing everything in entirety

How to Deploy the Assembly Line of AI for Discoveries: First and foremost, know what you are looking for. Once you do, break it down into the key steps required to solve the problem. The complexity of most discoveries means that you will have control over or be working on only one bit. Find that one bit, gather all the information about it, and input it into the relevant AI system for data and information processing.

Once you get the appropriate results, test them and run them mathematically and practically to see if they make sense. If they do, write a paper and let others comment on it while you continue your research.

Once feedback is gathered, process it to create a new, high-quality dataset for your needs. Use this to train your AI model. Once trained, your AI can start churning out data and insights that you can rigorously test to see if they align with practical tests. This will help you not just fine-tune but also lead to finding what you were seeking in the first place.

The widespread use of AI in scientific discoveries will make the year 2024, by 2124, appear as pivotal as 1924 does to us today—a stepping stone into a remarkable era of scientific and technological advancement. In hindsight, we may admit that the assembly line for science truly began during these years.

In case you missed:

Satyen is an award-winning scriptwriter, journalist based in Mumbai. He loves to let his pen roam the intersection of artificial intelligence, consciousness, and quantum mechanics. His written words have appeared in many Indian and foreign publications.

Leave A Reply

Share.
© Copyright Sify Technologies Ltd, 1998-2022. All rights reserved