Why agricultural advice still needs science, evidence and field judgment

By: Fayaz Ahmad

We are living in what Zygmunt Bauman called “liquid modernity”—a condition in which institutions, relationships and social structures are increasingly fluid. I would suggest that knowledge, too, has become liquid. What is regarded as established knowledge today can be revised, refined or even overturned tomorrow by new evidence.

Thomas Kuhn’s conception of the paradigm shift captured those moments when accumulated anomalies and new discoveries compelled a fundamental change in how a field understood itself. Such transformations were relatively infrequent. The information revolution has altered that rhythm. Knowledge is now produced, circulated, challenged and revised at unprecedented speed.

From Max Weber’s rationalisation to Daniel Bell’s post-industrial society and Alain Touraine’s programmed society, sociologists have attempted to understand how transformations in knowledge and technology reshape social organisation. Artificial Intelligence (hereafter AI) may be taking us into another phase. I have described this possibility as an “unprogrammed society”—a condition in which increasingly autonomous technologies participate in the production, interpretation and circulation of knowledge.

AI is already transforming medicine, astronomy, scientific research, engineering and education. In agriculture and horticulture, it is being used for weather forecasting, disease and pest prediction, crop monitoring, image-based diagnosis, irrigation management and yield prediction.

The possibilities are enormous.

But so is the philosophical question that accompanies them:

Does democratising access to information democratise expertise?

I believe the answer is no.

The illusion of mastery

This is where a concept I have developed over the years—“maleducation”—becomes relevant. I do not use the term to mean illiteracy or lack of schooling. Rather, I use it to describe a distortion of education in which literacy, qualifications and access to information produce an illusion of mastery.

A person may be able to read about almost anything, search Google, consult AI and collect enormous quantities of information. Yet the ability to access information does not necessarily provide the ability to understand its context, evaluate its evidence, recognise its limitations or apply it appropriately.

Information can therefore increase while understanding remains shallow.

Confidence can grow faster than competence.

And, paradoxically, greater access to knowledge can sometimes produce less intellectual humility.

When AI becomes an authority

AI can amplify this tendency.

This is particularly visible in agriculture, where farmers, pesticide dealers and crop advisors increasingly use AI to identify diseases, select fungicides and formulate crop-protection recommendations.

There is nothing inherently problematic about this. AI can be an extraordinarily useful assistant.

The problem begins when an AI-generated answer becomes an authority rather than a hypothesis to be verified.

I recently experienced this while reviewing the management of sooty blotch and flyspeck (SBFS) in apple. I spent several hours examining research papers and recommendations from scientific sources, including the American Phytopathological Society, ScienceDirect and University of Illinois Extension.

I subsequently cross-checked the information through several AI applications. The responses were impressive. They were rapid, detailed and often highly informative.

But some recommendations did not correspond with findings reported in the scientific literature.

When I presented the conflicting evidence to the AI chat systems and asked them to explain the discrepancy, they revised their responses and, in some cases, acknowledged their earlier error.

That raised a disturbing question:

What if I had not encountered the scientific report that contradicted the AI recommendation?

As a farmer advisor, I could have passed inappropriate advice to a grower. A farmer could then have spent money on a product, invested labour in its application and still failed to control the disease.

The cost of such an error is not merely financial.

For a farmer, an ineffective intervention can mean lost time, continued disease development, uncertainty and psychological distress.

AI does not eliminate expertise

This experience reinforced something that applies far beyond agriculture.

AI does not eliminate the need for expertise. It changes the nature of our responsibility towards expertise.

We should neither worship expertise nor dismiss it.

Scientists, researchers and extension institutions can be wrong; science advances precisely because established claims are questioned. But scientific questioning requires evidence, methodology and engagement with the existing body of knowledge.

There is an important difference between saying, “This recommendation is wrong because new evidence demonstrates otherwise,” and saying, “This recommendation is wrong because an AI application gave me a different answer.”

The first is scientific inquiry.

The second may simply be an illusion of expertise.

From digital literacy to epistemic literacy

This is why the future citizens, professionals and agricultural advisors will need more than digital literacy.

We will need epistemic literacy—the ability to ask where a claim comes from, what evidence supports it, how reliable that evidence is, under what conditions it applies and what its limitations are.

AI can dramatically increase our access to knowledge.

But access is not understanding, information is not expertise, and confidence is not competence.

The challenge before us is therefore not whether we should use AI.

We should.

The challenge is whether we can use it without surrendering our capacity to question it.

Use AI. Read the research. Consult the expert. Observe the field. Verify the evidence. And only then advise.

The real danger of the information age may not be ignorance.

It may be the illusion that because we can find an answer, we have understood it.

Dr. Fayaz Ahmad Bhat is a former academician with a background in Sociology and postgraduate teaching and research. He is presently engaged as a farmer advisor, working closely with farmers and agricultural stakeholders in the field of horticulture and crop protection. His interests include agriculture, plant protection, farmer advisory systems, sociology of knowledge and the changing relationship between technology and society.