“When the capabilities of artificial intelligence (AI) surpass those of human beings, we will face two outcomes: immortality and death.” – Nick. Bostrom, AI ethics expert/philosopher
On November 17, 2021, Science announced the top ten scientific breakthroughs of the year, and the artificial intelligence AlphaFold ranked first in the prediction of protein structure. It predicts protein structures with atomic-level accuracy and can predict 98.5% of human protein structures. The human protein structure information provided by this database is more than twice the amount of information accumulated by humans so far, which is not only accurate, large in quantity, but also fast.
With the increasing application of AI in the field of medicine, a series of controversies have followed, and people have more and more questions about AI.
1. At this stage, is artificial intelligence in the medical field really AI?
2. In drug research and development, where can AI be applied and to what extent has it progressed?
3. Will AI replace traditional drug research and development in the future, and where should companies and practitioners go?
What exactly is artificial intelligence (AI)?
In 1956, scientists such as John McCarthy first proposed the concept of artificial intelligence (AI) at a seminar at Dartmouth College in the United States, marking the birth of artificial intelligence. The definition they gave at the time was this: Artificial intelligence is the science and engineering of making intelligent machines, especially intelligent computer programs. We can simply understand that AI is a program, an intelligent computer program written by humans. With the development of society and the advancement of science and technology, people’s demand for artificial intelligence is increasing, and the cognition has deepened, and its definition has been further expanded. In an article “History, Status and Future of Artificial Intelligence” published by the Chinese Academy of Sciences in 2019, scientists explained artificial intelligence as follows:
“Artificial intelligence is a new technical science that researches and develops theories, methods, technologies and application systems that can simulate, extend and expand human intelligence. The purpose of research is to promote intelligent machines to listen (speech recognition, machine translation, etc.) and see (image recognition, text recognition, etc.), can speak (speech synthesis, human-machine dialogue, etc.), can think (human-machine game, theorem proving, etc.), can learn (machine learning, knowledge representation, etc.), can act (robot, automatic driving a car, etc.).”
Today, AI is not just limited to a computer program, it is also an all-encompassing composite science that is applied to various industries, including the field of medicine. Taking medical diagnosis as an example, CARE, a Class III medical device product approved by NMPA in 2021, is an ophthalmic artificial intelligence software for diagnosing fundus diseases. The overall accuracy was improved from 0.921 to 0.952. This product was developed by a research team at Sun Yat-Sen University, and the research results were published in The Lancet Digital Health.
As early as the 1960s and 1970s, foreign AI application research in the field of medicine began. In 1968, the world’s first AI system DENDRAL for inferring chemical molecular structures developed by Edward Feigenbaum, a computer scientist at Stanford University, came out. In the 1970s, Edward H. Shortliffe of Stanford University and others developed the first AI medical system MYCIN for the diagnosis, treatment and consultation of blood infections.
So the first question already has an answer. Is artificial intelligence at the current stage of the medical field considered AI in the true sense? Yes.
2 What fields has AI drug R&D been applied to?
The entire process of drug development can be simply summarized as: theoretical research, drug discovery, preclinical research, clinical research, and other stages.
A 2021 “White Paper on AI-Assisted Drug R&D Actions of the World’s 44 Top Pharmaceutical Companies” shows that more than two-thirds of AI applications are concentrated in the drug discovery stage, including target determination, biomarker selection, lead compound determination, and structure. The study of the effect relationship, the screening of active compounds, and the selection of candidate drugs, etc. Followed by the clinical treatment stage, accounting for about a quarter, including disease diagnosis, precision medicine, treatment outcome prediction, data analysis, pathological research, drug compliance and new therapy development, etc.
From the perspective of disease treatment areas involved, AI drug research and development are mainly concentrated in the three major disease areas of cancer, mental health and cardiovascular disease, which are also matched with disease prevalence, clinical needs and market capacity.
Among the participants are many well-known MNCs such as Pfizer and Novartis, as well as experienced Internet giants Google, Microsoft, etc., such as the AlphaFold mentioned at the beginning, from Google, as well as some scientific research institutes and AI startups.
At present, one of the fastest-growing AI drugs in the world, Igalmi® (dexmedetomidine, dexmedetomidine), has been launched. It is a sublingual film for agitation episodes in patients with bipolar disorder or schizophrenia, developed by BioXcel Therapeutics in the United States and approved by the FDA on April 6, 2022. BioXcel is a clinical-stage biopharmaceutical company incorporated in Connecticut, USA in 2017. They combine big data and proprietary machine learning algorithms to screen products from existing approved drugs or clinically validated drugs, identify new ones The therapeutic index of neuroscience, immuno-oncology drug research and development. The company’s two fastest-growing projects are BXCL501 and BXCL701. BXCL501 is the marketed Igalmi®, while BXCL701 is still in Phase I/II clinical phase. It is an investigational, orally administered systemic innate immune activator. For the treatment of relapsed or immune checkpoint inhibitor-naïve aggressive prostate cancer and advanced solid tumors.
Then the second question: in which areas of drug research and development, AI has been applied, and to what extent has it progressed?
AI is applied in all stages of drug research and development. Currently, it is mainly concentrated in the two stages of drug discovery and clinical research. The diseases are mainly in the fields of cancer, psychiatry and cardiovascular. The fastest-growing AI drugs have been launched, and many are still in the early development stage. The fastest-growing AI drugs in China have reached clinical phase I. It can also be seen that the global AI drug research and development still belongs to weak artificial intelligence, not strong artificial intelligence, nor super artificial intelligence. In other words, the current stage of AI drug research and development is only to assist human work. In practical applications, they only play a role in one aspect, only to solve the problem of a specific drug development stage. Their creative ability, intelligence level, etc., have not yet reached the level of completely replacing human beings, or even surpassing human beings.
3. The ultimate question: Will AI replace traditional drug research and development in the future?
The director of Stanford University’s Artificial Intelligence Laboratory, Chinese-American professor Wu Enda, once said: “The progress of science and technology is not to replace human beings, but to release human beings from repetitive labor to do more meaningful things.” AI drug development is not a replacement Traditional drug research and development is to free human productivity from traditional and inefficient work, let human beings do things that require more wisdom and creativity, and create more and new employment opportunities for human beings. So, what should companies and practitioners do?
No matter how technology advances and industries change, the only constant core is innovation. This is not only the foundation of the pharmaceutical industry, but also the foundation of all enterprises. For each of us practitioners, it is to always maintain a curiosity about the world, to constantly grow and update, to constantly iterate on ourselves, and to make ourselves a lifelong learner. Sheikh Zaki Yamani, former Saudi oil minister, once said: “The Stone Age did not end because there were no more stones, and the oil age did not end because there was no more oil.” Will AI end the era of traditional pharmaceuticals? What is your answer?
