This article aims to draw parallels between vast scientific discoveries gained with the writings of Alchemist Paracelsus and Artificial Intelligence (AI) for modeling and predictive analytics
Alchemist Paracelsus, your cough medicine, and toxicity
Without thinking, you read (or should read) the dosage instructions on the back of your cough medicine bottle before you take a dose to relieve the irritating throat tickle from your cold. You drift off to sleep, feeling better, with the effective ingredients in the cough syrup doing their job. We have Renaissance alchemist and physician Paracelsus to thank for those dosing instructions. We owe him a debt of gratitude for creating the standard of effective ingredients in appropriate proportions targeted to our symptoms.
Concepts of dosing, proportions, and symptom or organ targeting, detailed in the writings of Paracelsus, lead to a lower chance of toxicity (harm).
Paracelsus, born in 1493 as Theophrastus Bombastus Von Hohenheim, practiced science, alchemy, and medicine during the Renaissance. Many in medicine and science consider Paracelsus the most influential physician and scientist in Western Europe during the Renaissance. He passed in 1541, sadly, while still not accepted by physicians and academics of his day.
Paracelsus wrote, “What is there that is not a poison? All things are poison, and nothing is without poison. Solely the dose determines that a thing is not a poison.” Medicine, toxicology, and pharmacy have pivoted on this profound statement. Paracelsus established the difference between therapeutic (helpful) and toxic. He is now considered the Father of both Toxicology and Pharmacology; both areas of study have saved untold lives.
We generally understand that if we take too much cough syrup or a pain reliever, we may experience toxic effects. In Western Europe, people did not understand toxicity related to the amount taken in Paracelsus’ time and before.
Paracelsus and his use of Scientific Methodology
“Nature hints at cures,” wrote Paracelsus, regarding his belief that herbal remedies did not have the potency to treat many afflictions. He delved into the world of inorganic salts, metals, and minerals to treat diseases and discomforts.
Paracelsus used the science of experimentation to prove theories gained from observation. In this methodology, he documented evidence-based knowledge.
His fellow physicians were horrified by his writings and unwillingness to accept the medical treatment theories of the day: primarily that blood, bile, phlegm, or the humors determined health.
He often had contentious discussions and relationships with his peers. In his conviction that he had the correct approach, he was often blunt with fellow academics. His works were often banned, as they argued against the accepted theories of the day.
After his death, Paracelsus’s writings so deeply reformed medicine the academic and medical communities often refer to him as the Luther of medicine.
Through meticulous dosing experiments, Paracelsus successfully utilized mercury to treat syphilis.
He observed and wrote about the toxicities suffered by the miners of the Tyrolean Mountains from inhaling the dust and fumes in the mines and others who processed the metals brought up from the mines. Observing the illnesses experienced by the workers’ families, Paracelsus connected the miners’ family members’ exposure to the metal dust that clung to the miners’ clothing and the maladies the family members experienced. He advocated for ventilation and protection for the miners and workers. Sadly, not many who could make a difference paid attention.
With scientific methodology, Paracelsus developed glass, enamels, imitation metal, and gems practicing what we now call industrial chemistry. Some of his notable actions, writings, and accomplishments include:
- introduction of dosing, the systematic use of chemical substances in proportions, measurements, and time increments for relief of discomfort and to treat disease with the least toxic effect
- writing the first known occupational medicine book
- advocating for the cleaning and draining of gunshot wounds; differing from the standard treatment of the day, which consisted of packing wounds with peat and lining them with balms and salve, preventing drainage
- writing that illnesses were related to specific organs and identified chemicals had a more significant effect than others on treating diseases associated with a particular organ
- development of the theory of targeted organ toxicity
- establishment of scientific methodology in medicine and chemistry
You may be inclined to give Paracelsus a nod of gratitude with the next couple of tablets you take for headaches, allergies, cough, or health issues. Go for it. His brilliance and tenacity have benefited us all.
Enlightenment for chemistry took an overly long time. Prejudice against the science of chemicals and their interactions continued, and chemistry did not gain recognition as a legitimate science until the late 1800s. In modern times, that lack of recognition by the scientific community is unfathomable.
The complexity of problem-solving takes us beyond experimentation to Artificla Intelligence with computer modeling and simulation in Scientific Discovery
Our timeline takes us from the late 1800s to an entire century later. Until the mid-twentieth century, there were two sciences: experimental and theoretical. For the experimental arm of science, we can thank Paracelsus.
In the mid-Twentieth century, many problems proved too complicated to be solved with experimentation and observation. Researchers turned to computers to solve complex problems. Computer modeling and simulations opened a whole new world of discovery and problem-solving. The machines used for these simulations became known as intelligent machines or artificial intelligence (AI).
Neural Networks and the Needle in the Haystack
We have accumulated a vast array of knowledge, a knowledge explosion. Mining this knowledge, data mining, for relevant information and not wanting to miss that potentially critical needle in the haystack has become imperative.
Intelligent machines, Artificial Intelligence (AI), are the current answer to that problem.
- AI functions to keep research from being cost-prohibitive
- AI gives us new theories to test and problems to solve, problems we did not we had until AI identified them (the “you don’t know what you don’t know” conundrum)
- AI won’t replace human thinking or creativity; AI thinks differently from humans, making it a helpful tool.
Some intelligent machines have multi-layered Neural Networks (NN), making them good at discovering variances and patterns in ginormous mountains of information. Agrawal’s Needle in the Haystack gives us insight into how NN intelligent machines work, using scientific search and discovery approaches. NN AI’s take these approaches to a new level of thought and analyses beyond human researchers’ grasp while mining through the mountain of information.
Per Agrawal, intelligent prediction machines:
- take the search function and arrange the haystack without physically rearranging the haystack; human researchers would have to tear apart the haystack to find the needle
- use the discovery function to find the variance in the pattern from hay and locate the needle
Modern intelligent machines are described by (Seeber et al.) as teammates which:
- define problems
- identify root causes
- propose and evaluate solutions
- choose between different options
- make plans
- take actions
- learn from interactions
Thus they are intelligent machines, or Artificial Intelligence (AI), which possess General Purpose Technology GPT. Did you have an “aha” moment for the acronym “GPT,” as I did?
Now, you know what GPT stands for, which can make for a great water cooler or happy hour conversation!
Back to our AI (but a good Gin and Tonic sound divine to me right now) GPT, and the search and discovery mission of finding the needle:
- AI allows researchers to effectively wrangle the knowledge explosion to enable scientific discovery with efficiency
- AI’s ability to rapidly find the needle in the haystack that researchers were not even sure was there makes it an excellent teammate for research
The problem-solving ability of AI is vast. It ensures humans don’t miss a relevant piece of information in their research. It takes a truly deep dive holistic view of the problem and the involved elements and finds the needle in the haystacks which researchers (yes, humans) identify as key to the resolution of the problem. It does not make mistakes due to sleep deprivation or too many Gin and Tonics the prior evening during happy hour with friends.
Three ways in which Artificial Intelligence (AI) supports scientific discovery
As described by Krenn et al., 2022 there are three ways in which AI supports scientific discovery:
- AI can provide information through a computational microscope. This
- occurring in simultaneous multi-layered and directional ways that humans cannot imitate
- through the computational microscope, AI creates new data that human scientists use to further their understanding.
- AI provides unique resources for information and ideation, greatly expanding human creativity and idea-formation
- AI looks for new ideas or unexpected connections (“you don’t know what you don’t know”)
- human scientists use these further expand their understanding
- AI acts as a catalyst for enlightenment, an ‘agent of understanding’ (Krenn et al.)
- replacing the human in making conclusions
- transferring these new scientific concepts to different phenomena
Real-world Artificial Intelligence (AI) applications to boost discovery, imagination, and thinking
Today, AI is used to improve our world, decrease risk, improve medical care, and countless other uses. Some examples are:
Micro and craft beer brewers use AI to assist them in dealing creatively with supply chain issues, find great flavor combinations, and increase customer engagement.
AI is assisting the World Bee Project in examining why the world’s bee population is declining and what we can do to save our bees. Without AI, the data analyses of vast amounts of collected bee information would be hugely expensive, take a long time, and is herculean without the aid of AI.
Earthquakes will cause less disruption to the infrastructure of the city of Los Angeles as AI assists in prioritizing upgrades to the water supply system through computer modeling of various scenarios. With AI as a team member, critical institutions like hospitals have the least chance of failing water lines with a major earthquake.
A healthcare example, which may please Paracelsus, AI is being put to work as a radiologist’s assistant, augmenting the review and reading of images performed by radiologists. From my personal experience of sitting on tumor boards, I know there is often much debate about what a diagnostic image may be depicting.
AI and Predictive Analytics
AI can take the needle in the haystack a step further by running predictive analytics, where it will take:
- historical data
- computer modeling data
- statistics
- complex and intelligent machine processes to identify risks and opportunities
An evidence-based example of the power of predictive analytics exists in its use to identify vulnerable neonatal ICU infants at risk for deterioration in their condition, enabling NICU caregivers the opportunity to intervene
- Predictive analytics examines and models scenarios based on several data points
- The points of data are individual NICU infant assessment points documented by a multitude of staff from various disciplines over a period of time
- AI predictive analytics models possible deterioration from all of these data points on a consistent basis, more effectively and rapidly than a person is able
Calls for caution
There are calls for caution in the development of AI. In March 2023, several giants of the technology industry, scholars, artists, and others published an open letter through the “Pause Giant AI Experiments: An Open Letter,” demanding a six-month pause in the development of more powerful than ChatGPT4 until guarantees are put into place that AI is used for positive advancements and safeguards.
Full circle from Artificial Intelligence to Paracelsus: Scientific Discovery
The knowledge explosion is possible because Paracelsus did not bend to his peers. He remained steadfast in his belief in experience, observation, and experimentation, collecting cold, hard data. He placed a greater value on and saw it as more ethical to advance knowledge with experimentation and the associated data.
Both Paracelsus and AI established legitimate and groundbreaking paths for us to examine problems and find their solutions through scientific discovery.
Works Cited
Bianchini, Stefano, et al. “Artificial Intelligence in Science: An Emerging General Method of Invention.” Research Policy, vol. 51, no. 10, Dec. 2022, doi:10.1016/j.respol.2022.104604.
Borzelleca. “Paracelsus: Herald of Modern Toxicology.” Toxicological Sciences, vol. 53, no. 1, Jan. 2000, pp. 2–4, doi:10.1093/toxsci/53.1.2.
Crowell, Chris. “How Deep Liquid’s AI Improves Beer Recipes and Engages Customers.” Craft Brewing Business, 15 June 2022, https://www.craftbrewingbusiness.com/featured/how-deep-liquids-ai-improves-beer-recipes-and-engages-customers/.
Gravenstein. “Paracelsus and His Contributions to Anesthesia.” Anesthesiology, vol. 26, no. 6, Nov. 1965, pp. 805–11, doi:10.1097/00000542-196511000-00016.
Gravenstein, J. S., MD. “Paracelsus, Five Hundred Years: Three American Exhibits.” ASA Publications, American Society of Anesthesiologists, Inc, Nov. 1965, https://www.nlm.nih.gov/exhibition/paracelsus/aftermath.html.
Hajdu, Steven I. “Two Pioneering Chemists, Three Hundred Years Apart.” Annals of Clinical & Laboratory Science, vol. 35, no. 1, Jan. 2005, pp. 105–07.
Krenn, Mario, et al. “On Scientific Understanding with Artificial Intelligence.” Nature Reviews Physics, vol. 4, no. 12, Oct. 2022, pp. 761–69, doi:10.1038/s42254-022-00518-3.
Marr, Bernard. “10 Wonderful Examples of Using Artificial Intelligence (AI) for Good.” Forbes, 22 June 2020, https://www.forbes.com/sites/bernardmarr/2020/06/22/10-wonderful-examples-of-using-artificial-intelligence-ai-for-good/?sh=47704ce12f95.
Nicholas, Spyros, et al. “Theophrastus Bombastus Von Hohenheim (Paracelsus) (1493–1541): The Eminent Physician and Pioneer of Toxicology.” Toxicology Reports, vol. 8, Feb. 2021, pp. 411–14, doi:10.1016/j.toxrep.2021.02.012.
Pedrick, Alexis, and Lisa Berry Drago. Alchemical Origins of Occupational Medicine: From Paracelsus to OSHA. Science History Institue, 25 Aug. 2020, https://www.sciencehistory.org/distillations/podcast/the-alchemical-origins-of-occupational-medicine.
ProCon.org Encylopedia Britannica. “Is Artificial Intelligence Good for Society? Top 3 Pros and Cons.” ProCon.Org, edited by Natalie Leppard, 30 Mar. 2023, https://www.procon.org/headlines/artificial-intelligence-ai-top-3-pros-and-cons/#59.
Seeber, Isabella, et al. “Machines as Teammates: A Research Agenda on AI in Team Collaboration.” Information & Management, vol. 57, no. 2, Mar. 2020, doi:10.1016/j.im.2019.103174.
“What Is Predictive Analytics?” IBM, https://www.ibm.com/topics/predictive-analytics. Accessed 15 May 2023.
Moorman, J. Randall. “The Principles of Whole-Hospital Predictive Analytics Monitoring for Clinical Medicine Originated in the Neonatal ICU.” Npj Digital Medicine, vol. 5, no. 1, Mar. 2022, pp. 1–6, doi:10.1038/s41746-022-00584-y.
Leave a Reply