You want the latest insights on artificial intelligence, delivered第一时间.
Reprinted from Deep Science Review, for academic sharing only. If there is any copyright infringement, please leave a message to request removal.
Editor's Note
In an era where AI is profoundly reshaping the paradigm of scientific research, the more powerful technology becomes, the more precious human judgment, curiosity, and collaborative spirit become. AI can design proteins, but it is always humans who ask the right questions and give meaning to the answers.
Many have questioned whether AI will eventually seize human jobs and escape human control. Recently, Nobel Laureate David Baker provided a clear answer: "AI is, after all, just a tool."
Compiled by Lin Yan
AI is transforming the efficiency and methods of scientific research—computations are faster, predictions are more accurate, and designs are more complex. The progress of AI is overwhelming.
However, the more powerful the technology, the clearer the questions become: "Will AI replace researchers?" and "If so, where should researchers go from here?"
David Baker
In a recent interview, David Baker, the 2024 Nobel Prize winner in Chemistry and professor at the University of Washington, offered clear answers that alleviate researchers' anxieties.
He stated, "AI has completely revolutionized protein design research," but added, "AI is, after all, just a tool."
Artificial intelligence (AI) can not only predict protein structures but also design novel proteins with desired functions.
Professor Baker has utilized AI to design entirely new proteins that do not exist in nature, expanding the boundaries of life sciences. He continues his research with the goal of using AI to design proteins for developing and commercializing new drugs.
The following is the transcript of the interview:
01: AI is a Powerful Tool, But Humans Decide Which Questions to Ask
Question: What proportion of protein design research is currently driven by AI?
Baker: Our laboratory began seriously focusing on developing AI-based methodologies around 2018-2019, and we are still building the next generation of AI to design more complex proteins. From the perspective of methodology development, it is no exaggeration to say that almost all of our work over the past six years has utilized AI.
Question: Will AI replace researchers?
Baker: No. AI is just a very powerful tool. In the past, gene sequencing itself was a major research project; now, it is routine analysis—you put in a sample and get results the next day. I believe AI will eventually become that kind of tool as well.
Question: What are the limitations of AI?
Baker: What is more difficult than using AI to design proteins and verifying them in the lab is translating them into actual drugs and completing clinical trials. At this stage, public data is insufficient, and our understanding of biology and medicine remains incomplete. It is difficult for AI to achieve easy breakthroughs in this field. While it can assist in designing proteins better suited for production processes, it is too early to expect it to significantly shorten clinical development cycles.
Question: As AI becomes increasingly proficient at protein design, what capabilities will human researchers need?
Baker: The key remains asking questions. Determining which questions are important, planning how to verify design results, and assessing the extent to which AI outputs are trustworthy—these are the core competencies. The fundamental questions of science have never changed.
02: Talent is the Most Important Factor
Question: In the fields of AI biology and protein design, among computing power, talent, experimental facilities, and data, what is the most important?
Baker: Talent is the most important. Competitiveness ultimately comes from securing excellent researchers and supporting them to delve deeply into the problems they are interested in over the long term. Protein design is not a field that can grow in isolation. Basic sciences, biology, chemistry, and computation must all be strong together. We should support a diverse research foundation, upon which protein design can develop.
Question: What advice do you have for young researchers and graduate students?
Baker: Do not over-calculate the future; instead, seize the problems that interest you most and ignite your passion right now. New paths in science are often not walked out according to a predetermined plan but are naturally opened up during the process of deeply studying problems that interest you. I myself started with an interest in protein folding and structure prediction, which eventually led me to protein design. What truly matters is thinking deeply and for a long time about the issues you genuinely care about, and working with like-minded people.
03: Future Interests Include Nanomachines and Agriculture
Question: How do you create an environment where researchers can collaborate freely and delve deeply?
Baker: I value an environment where people naturally gather and communicate continuously. Our laboratory provides free food every day; anyone can come, eat, and chat. We also share everything without reservation. I believe difficult problems are better solved through collaboration than by working alone. I call it a 'collective brain'—every researcher should connect and interact like neurons. This is also why I stay in the laboratory almost all the time. Only this way can I immediately see who is encountering difficulties or where a breakthrough has occurred, and offer help.
Question: After winning the Nobel Prize, you must have received many speaking invitations and external requests. How do you maintain balance?
Baker: I rejected most of the invitations. I still spend a large amount of time in the laboratory, just as I did in the past.
Question: What new fields do you want to explore in the next 5 to 10 years?
Baker: It is hard to predict accurately. However, I see huge potential in nanomachines—ultra-micro mechanical systems that perform specific functions at the molecular or protein level. They have broad application prospects not only in medicine but also in technology. Agriculture is also very interesting. With global warming, we can use protein design to create plants that are more stable at higher temperatures. The only certainty is that I will not be doing the same research as I am now.
Question: Why do you constantly seek new research topics?
Baker: I get bored relatively easily. I do not like repeating the same work. When a topic catches my interest, I will hold onto it, think about it, and then naturally move into new areas that others have not yet touched.
Question: What is the ultimate goal as a scientist?
Baker: There is no specific ultimate goal. I enjoy the process of discovering new things in the laboratory with graduate students and postdocs. Of course, there are many major issues to solve, such as neurodegenerative diseases, but rather than setting a single goal, I prefer the process of continuously tackling important problems itself.
Original Article Link:
Baker says AI designs proteins, but people set questions and drive research
https://biz.chosun.com/en/en-science/2026/04/08/Y2HAKUH2XNHBXBDHEXUKCPAUUA/