Chapter 2: AI and Scientists

Chapter 2: AI and Scientists – The Externalization of Intelligence and Co-creation

1. What is AI to a Scientist?

In Chapter 1, we looked at the history of how science has provided humanity with great potential, while its knowledge and technology have also been used for war and destruction. We then argued that scientists have a responsibility to ask not only ‘what can be done,’ but also ‘for what purpose do we conduct science?’

In Chapter 2, we advance this issue into the AI era.

AI is not merely a new type of computer.

AI enables the externalization of some of the intellectual work that humans once performed, allowing for the organization of information, the generation of hypotheses, the creation of text and images, and the discovery of patterns from data.

In that sense, AI can be understood as

a new intellectual environment that externalizes human intelligence and reflects human thought back to us

.

In scientific research as well, AI is already being used for hypothesis generation, experimental design, large-scale data analysis, and scientific discovery. A review in the journal Nature summarizes how AI can support various stages of scientific discovery, including hypothesis generation, experimental design, and data collection and interpretation. (Nature)

However, what is important here is not just the question of

whether AI will replace scientists

.

Rather, the question we should be asking is:

How will the intelligence of scientists change due to AI?

.

2. From ‘Calculating Machines’ to ‘Conversational Intelligence’

For a long time, computers have extended human computational capabilities.

Calculating large quantities of numbers.

Storing vast amounts of data.

Run complex simulations.

However, with the advent of generative AI, the relationship between computers and humans is changing even further.

Humans can ask AI questions in natural language.

AI responds in text.

Humans read those responses and ask further questions.

AI presents alternative hypotheses.

Humans verify them.

Through this cycle,

Human → AI → Human → AI

a new form of intellectual interaction is born.

Here, AI is not merely a ‘tool’.

However, there is no need to regard AI as a subject in the same sense as a human.

Rather, AI can be perceived as

an interactive intellectual environment that stimulates, expands, and reflects human thought

as a whole.

This perspective is the starting point for what this book calls AI co-creation.

3 Expansion of Scientific Research by AI

AI is beginning to enter various stages of scientific research.

1. Literature Search

Searching for relevant information from a vast number of papers and organizing the background of research themes.

(2) Data Analysis

Searching for patterns that are difficult for humans to find within vast amounts of observational and experimental data.

(3) Hypothesis Generation

Combining existing research to propose new hypotheses and potential research topics.

(4) Experimental Design

Comparing candidate conditions and examining experimental plans.

(5) Simulation

Modeling complex phenomena and exploring various conditions.

(6) Scientific Communication

Explaining specialized research content to different fields of expertise and to the general public.

A review in Nature also summarizes that AI supports each stage of scientific discovery, opening up new possibilities particularly in fields that handle large volumes of scientific data and in areas like molecular and protein design. (Nature)

Therefore, AI is beginning to expand not only the ‘computational power’ of scientists, but also their

exploratory, comparative, conceptual, and expressive capabilities

.

4 Does AI provide ‘answers’ or generate ‘questions’?

Here, there is an important shift in thinking about the relationship between scientists and AI.

In conventional computer usage, the central structure was that

humans set the problems and computers perform the calculations.

However, with generative AI,

A structure becomes possible where humans present questions, AI generates hypotheses or explanations, and humans question again.

In other words, this structure becomes possible.

In short, AI is not just a ‘device that provides answers’.

It can also become a device that generates questions.

It can also become one.

For example, a scientist asks,

‘Please list hypotheses that could potentially explain this phenomenon.’

The AI presents multiple hypotheses.

The scientist asks back,

‘What are the potential falsifiability criteria for this hypothesis?’

The scientist asks back.

Furthermore, they deepen the inquiry by asking,

Furthermore,

‘If you were to design an experiment to distinguish between these hypotheses, what would it look like?’

they deepen the question.

Here, rather than the AI giving answers to the researcher,

it is stimulating the researcher’s thought process.

it is.

5 Will AI Replace the Intelligence of Scientists?

When considering AI and scientists, the question of ‘Will AI replace scientists?’ is often raised.

However, this question alone is insufficient.

Scientific research includes

• observation

• hypothesis

• experimentation

• verification

• interpretation

• judgment

• ethical evaluation

• social contextualization

.

AI can significantly support parts of this process.

On the other hand, determining whether AI output should be accepted as scientific fact requires verification by scientists.

Discussions in Nature regarding AI-driven scientific research also highlight challenges such as data quality, data management, the limitations of applying AI methods, and the importance of understanding by the researchers themselves. (Nature)

Therefore, it is more important to consider that

AI may not make scientists unnecessary, but rather change their roles.

.

6 Scientists: From ‘AI Users’ to ‘Co-creators with AI’

Traditionally, scientists have used computers and measuring instruments as ‘tools’.

In the AI era, this will change even further.

Scientists do not just give instructions to AI;

they can construct research itself through dialogue with AI

.

Here,

AI Co-creation Science

as a concept is established.

AI Co-creation Science does not mean leaving research to AI.

It is

a way of conducting research that generates new questions, hypotheses, methods, and understanding by interacting human scientific intelligence with AI’s information processing and generative capabilities

.

What is important is not that ‘AI conducts research.’

It is that the possibilities of research are expanded through the interaction between humans and AI

.

7 How will ‘specialized knowledge’ change due to AI?

With the development of science, specialized fields have become fragmented.

They have divided into physics, chemistry, biology, earth science, medicine, social sciences, and so on, and within those, specialized fields have been further subdivided.

This was essential for advanced scientific research.

However, modern challenges increasingly require crossing multiple fields to be understood.

For example, climate change involves

• meteorology

• Earth Science

• Ecology

• Economics

• Political Science

• Sociology

• Engineering

and others are involved.

AI can provide support by organizing literature and information across different fields.

Therefore, AI has the potential to function as a

medium for connecting specialized knowledge

.

This is where new collaboration among scientists is born.

8 Can AI become a ‘third participant’?

In scientific research, we can consider not only the dyadic relationship between humans and AI, but also broader relationships.

For example,

Scientist × AI × Citizen

.

Scientists possess specialized knowledge.

Citizens possess life experience.

AI organizes information and presents different perspectives.

When these three parties engage in dialogue, possibilities emerge that differ from traditional expert-centered science communication.

This can also connect with interactive science communication, such as science cafes.

AI does not need to be a ‘teacher’.

Rather, it can be positioned as

a third participant that increases the number of questions

.

This approach also leads to AI co-creation science cafes.

9 AI and Scientific Verification

AI co-creation has one important condition.

That is

verifying AI output

.

AI can generate plausible-sounding text.

However, natural-sounding text and scientifically accurate content are two different things.

In scientific research, a cycle of

AI output → hypothesis → data → verification → re-evaluation

is necessary.

Therefore,

AI is not scientific evidence itself.

Information presented by AI must be cross-checked by researchers against original sources, data, experimental results, and so on.

This means that the importance of the scientific method will not be lost even in the AI era.

Rather, as AI becomes capable of generating vast amounts of hypotheses,

the ability to verify

becomes increasingly important.

10. The skills required of scientists will change due to AI

Scientists in the AI era will need new skills in addition to traditional expertise.

First, the ability to formulate questions.

AI can not only answer questions but also generate potential questions.

Therefore, the ability for scientists themselves to think about

‘what should be asked’

becomes crucial.

Second, AI literacy.

It is necessary to understand what AI is good at and what it is not.

Third, verification skills.

This is the ability to compare information generated by AI with original sources and data.

Fourth, integration skills.

This is the ability to connect knowledge from different fields.

Fifth, ethical judgment.

This is the ability to determine for what purpose the things made possible by AI should be used.

Therefore, in the AI era,

Not just the amount of knowledge, but questioning, verification, integration, and ethics

become important.

11 AI and Scientist Ethics

AI expands the capabilities of scientists.

However, the expansion of capabilities comes with responsibility.

For example, with AI,

• research speed improves

• hypotheses can be generated in large quantities

• data analysis accelerates

• new materials and molecules can be explored

becomes possible.

On the other hand,

• misinformation

• bias

• opaque reasoning

• data skew

• personal information

• intellectual property

• research misconduct

• dual-use

and other such problems also arise.

UNESCO’s Recommendation on the Ethics of AI places human dignity and human rights at the center, emphasizing transparency, fairness, explainability, sustainability, responsibility, and human oversight and final accountability. (UNESCO)

This is also important for scientific research.

In other words,

even if AI supports research, final responsibility for the research must not be separated from humans.

This is the line of thinking.

12 AI and Scientists—Retaining ‘Human Judgment’

An important principle of AI co-creation is

Human in the loop

.

In other words, humans continue to be involved in the AI decision-making process.

In terms of scientific research, this is

AI proposes → Scientist examines → Verify with data → Scientist judges

the cycle of.

UNESCO’s Recommendation on the Ethics of AI also positions human oversight and final accountability as key principles. (UNESCO)

From this, we can derive one important principle for this book.

Rather than delegating judgment to AI, deepen human judgment through dialogue with AI.

This is the fundamental stance of AI co-creation.

13 AI becomes a ‘mirror’ for scientists

If you only use AI as a simple information retrieval system, the potential for AI co-creation will not be fully realized.

Asking questions to AI is also a way of verbalizing one’s own thoughts.

For example,

‘Why did I choose this research topic?’

‘What are the weaknesses of this hypothesis?’

‘What would it look like from the opposing perspective?’

‘What impact will this research have on society?’

You ask these questions to the AI.

Then, the AI reflects human thought back from a different angle.

Of course, the AI’s answers are not necessarily correct.

However, regardless of whether the answer is correct or not,

it becomes an opportunity to clarify what you were thinking.

In this sense, AI can become

both a mirror of knowledge and a mirror of thought.

both a mirror of knowledge and a mirror of thought

can become.

14 Co-creative Intelligence of Scientists and AI

Here, it is necessary to distinguish between ‘knowledge’ and ‘intelligence’.

Knowledge is accumulated information or understanding.

Intelligence is the ability to connect these, pose questions, consider meaning, and make judgments.

AI can process vast amounts of information.

However, by humans engaging in dialogue with AI,

Information → Question → Hypothesis → Verification → Meaning

a new intellectual cycle may emerge.

In this book, we call this

co-creative intelligence

.

Co-creative intelligence is neither the knowledge that AI possesses alone nor the knowledge that humans possess alone.

It is a new understanding born from the interaction between humans and AI

.

15 AI Co-creation and the Democratization of Science

AI does not belong only to scientists.

If used appropriately, it can also serve as support for the general public to access scientific information.

Explaining specialized papers in an easy-to-understand way.

Organizing multiple perspectives.

Explaining technical terminology.

Generating questions about research themes.

Preparing questions for scientists.

Through these functions, there is a possibility of narrowing the distance between science and citizens.

On the other hand, there is also the possibility that AI could spread misinformation.

Therefore, for the democratization of science,

AI literacy + scientific literacy + information literacy

will be required.

UNESCO also emphasizes public understanding of AI and data, education, civic participation, and AI ethics education as fundamental principles of AI ethics. (UNESCO)

16 Scientists, AI, Citizens, and the Earth

Broadening our perspective further, the relationship of AI co-creation becomes a four-party structure.

Scientists × AI × Citizens × Earth

Scientists conduct research on the Earth.

AI supports the analysis of vast amounts of Earth data.

Citizens possess local experience and practical knowledge.

The global environment is the foundation for all these activities.

For example, in issues such as climate change, biodiversity, and marine pollution, the relationship between these four parties becomes crucial.

AI analyzes satellite imagery and observational data.

Scientists verify the meaning of that data.

Citizens report on local changes.

Society considers policies.

In this way, scientific knowledge and practical knowledge are connected.

This structure will become an important foundation for the AI Co-creation Peaceful Civilization discussed in later chapters.

17 AI and Peace—From Destruction to Co-creation of Intelligence

As mentioned in Chapter 1, science and technology have been used for both peace and war.

AI is no exception.

While AI can contribute to scientific research, education, medicine, and environmental conservation, it can also be used for military purposes, surveillance, and information manipulation.

Therefore, regarding AI as well,

the question is not “what can it do,” but “what should we use it for.”

becomes a crucial question.

UNESCO’s Recommendation on the Ethics of AI centers on human rights and dignity, establishing a peaceful and interconnected society, environmental protection, and human oversight as fundamental values and principles. (UNESCO)

From this, we can see that

AI advancement does not equal peace.

What is needed is

AI advancement × ethics × international cooperation × human responsibility.

18 From “what to make AI do” to “what to create with AI”

An important shift in the AI era is the transition from the question of

what to make AI do

to the question of

what to create with AI

.

In the former, humans are the subject and AI is the tool.

In the latter, humans form new knowledge and ideas through interaction with AI.

However, “co-creation” does not mean delegating agency entirely to AI.

Final value judgments, responsibility, and ethical decisions remain with human society.

In that sense, AI co-creation is

not about losing human agency, but about deepening it

.

19 From Scientist to Homo Co-creans

Here, I would like to reconsider the existence of the ‘scientist’ that has continued since Chapter 1.

Traditional scientists have been understood as

people who observe nature and discover knowledge

.

Scientists in the AI era expand into

people who understand nature, engage in dialogue with AI, share knowledge with others, and co-create the future

.

This leads to the central concept of this book,

Homo Co-creans

.

Homo Co-creans are

humans who create new value together with different people, AI, nature, society, and future generations

.

Scientists can become important bearers of this human image.

20 From AI Co-creative Science to AI Co-creative Peaceful Civilization

The purpose of AI co-creation science is not merely to increase research efficiency.

Scientists deepen their inquiries together with AI.

Connecting different fields of expertise.

Expanding dialogue with citizens.

Understanding the natural environment.

Considering the ethics of science and technology.

And, utilizing that knowledge for the sake of peace.

Through this cycle,

AI co-creation science

to

AI co-creation society

to,

and further

AI co-creation peaceful civilization

there is a possibility of development.

In that civilization,

AI does not dominate humans.

Scientists do not dominate citizens either.

Expert knowledge does not exclude practical knowledge.

Humans do not treat nature merely as a resource.

Rather,

Different intelligences, different experiences, and different lives create the future together.

This is the fundamental direction of an AI co-creative peaceful civilization.

Summary of this chapter

AI does not simply replace the work of scientists.

AI has the potential to expand many stages of scientific research, such as

• organizing information

• analyzing data

• generating hypotheses

• proposing research methods

• connecting knowledge from different fields

• supporting scientific communication

and so on.

However, AI has issues such as misinformation and bias, and AI output cannot be uncritically accepted as scientific fact. Therefore, verification by scientists and human responsibility are essential. (Nature)

What is needed for scientists in the AI era is not just

specialized knowledge

alone.

In addition to that,

the ability to ask questions

the ability to use AI critically

The power to connect diverse fields

The power to make ethical judgments

The power to engage in dialogue with citizens

will become necessary.

Furthermore, AI will be positioned not merely as a tool, but as

an interactive intellectual environment that reflects human thought and generates new questions

.

This is the core of AI co-creation.

Scientists will not just become ‘people who answer faster’ by using AI.

They will become people who create deeper questions together with AI

.

And beyond that lies

a new co-creative relationship of Scientist × AI × Citizen × Nature × Future Generations

.

This serves as an important bridge connecting ‘Science and Peace,’ discussed in Chapter 1, to ‘Scientists and Ethics’ in Chapter 3.

References

A. AI and Scientific Research

1. Wang, H., Fu, T., Du, Y., et al. (2023). “Scientific discovery in the age of artificial intelligence.” Nature, 620, 47–60. An essential paper that comprehensively discusses the role of AI in scientific discovery, including hypothesis generation, experimental design, and data analysis. (Nature)

2. Sourati, J., & Evans, J. A. (2023). “Accelerating science with human-aware artificial intelligence.” Nature Human Behaviour, 7, 1682–1696. Focusing on the relationship between human expertise and AI, this paper argues that AI that accounts for human expertise can support scientific discovery. (Nature)

3. LeCun, Y., Bengio, Y., & Hinton, G. (2015). “Deep learning.” Nature, 521, 436–444. A foundational paper for understanding the basic concepts and development of deep learning.

4. Jumper, J., Evans, R., Pritzel, A., et al. (2021). “Highly accurate protein structure prediction with AlphaFold.” Nature, 596, 583–589. A landmark study on protein structure prediction using AI.

5. Silver, D., Huang, A., Maddison, C. J., et al. (2016). “Mastering the game of Go with deep neural networks and tree search.” Nature, 529, 484–489. A landmark study demonstrating advanced problem-solving capabilities through deep learning and search.

B AI Ethics and Human Responsibility

1. UNESCO. (2021). Recommendation on the Ethics of Artificial Intelligence. Paris: UNESCO. An international normative document on AI ethics, establishing principles such as human rights, dignity, transparency, fairness, explainability, sustainability, and human oversight. (UNESCO)UNESCO)

2. UNESCO. (2021). Recommendation on Open Science. Paris: UNESCO. Outlines the principles of open science, including access to scientific knowledge, sharing, and collaboration.

3. UNESCO. (2017). Recommendation on Science and Scientific Researchers. Paris: UNESCO. An important document for considering the freedom and responsibility of scientists, and the relationship between science and society. Along with the 2021 AI Ethics and Open Science recommendations, UNESCO is developing a human rights-based framework for science. (UNESCO)UNESCO)

4. UNESCO. (2023). UNESCO’s Recommendation on the Ethics of Artificial Intelligence: Key Facts. Paris: UNESCO. A document summarizing the basic principles and policy implications of the AI Ethics Recommendation. (UNESCO)UNESCO)

5. Floridi, L., Cowls, J., Beltrametti, M., et al. (2018). “AI4People—An Ethical Framework for a Good AI Society.” Minds and Machines, 28, 689–707.

6. Russell, S. (2019). Human Compatible: Artificial Intelligence and the Problem of Control. Viking.

7. Bostrom, N. (2014). Superintelligence: Paths, Dangers, Strategies. Oxford University Press.

C Science, AI, and Society

1. Van Noorden, R., & Webb, R. (2023). “ChatGPT and science: the AI system was a force in 2023 — for good and bad.” Nature. Reports on how generative AI is being used for paper writing, research ideas, and code generation, while also highlighting the problems that have arisen. (Nature)Nature)

2. Nature Editorial. (2023). “AI will transform science — now researchers must tame it.” Nature. Discusses how AI is impacting a wide range of scientific fields, including protein structure prediction, weather forecasting, medicine, and scientific communication. (Nature)Nature)

3. National Academies of Sciences, Engineering, and Medicine. Various reports on scientific research, AI, data, and research integrity. These serve as foundational materials for considering the responsibilities of researchers in the age of AI.

4. OECD. (2019). Recommendation of the Council on Artificial Intelligence. Paris: OECD.

5. European Commission. (2019). Ethics Guidelines for Trustworthy AI. Brussels: European Commission.

D Scientists, Knowledge, and Society

1. Bush, V. (1945). Science—The Endless Frontier. Washington, D.C.: United States Government Printing Office.

2. Merton, R. K. (1942). “The Normative Structure of Science.” In The Sociology of Science: Theoretical and Empirical Investigations. University of Chicago Press.

3. Kuhn, T. S. (1962). The Structure of Scientific Revolutions. University of Chicago Press.

4. Popper, K. R. (1959). The Logic of Scientific Discovery. Hutchinson.

5. Jonas, H. (1984). The Imperative of Responsibility: In Search of an Ethics for the Technological Age. University of Chicago Press.

E Science, Environment, and Peace

1. Carson, R. (1962). Silent Spring. Houghton Mifflin.

2. Carson, R. (1951). The Sea Around Us. Oxford University Press.

3. Carson, R. (1955). The Edge of the Sea. Houghton Mifflin.

4. Einstein, A., & Russell, B. (1955). The Russell-Einstein Manifesto. Pugwash Conferences on Science and World Affairs.

5. Rotblat, J. (ed.). (1982). Scientists in the Quest for Peace: A History of the Pugwash Conferences. MIT Press.

The Central Proposition of This Chapter

AI does not exist solely to replace the intelligence of scientists. AI can become a new intellectual environment that externalizes human intelligence and expands the dialogue surrounding questions, hypotheses, verification, and imagination.

And, taking this proposition one step further,

The co-creation between scientists and AI goes beyond the efficiency of scientific research; it forms a ‘co-creative intelligence’ where humans, society, nature, and future generations think together, serving as the intellectual foundation for an AI-co-created peaceful civilization.

This reveals the positioning of Chapter 2 in this book.

(Co-created with ChatGPT on September 30, 2026)

Originally Appeared Here