AI in Healthcare and Agriculture

Two seemingly disparate fields — a provincial hospital’s X-ray room and a rice field in the Mekong Delta — are being transformed by the same type of technology: artificial intelligence that recognizes patterns from images and data. In both cases, AI does not replace humans but works as a tireless “second pair of eyes”: quickly scanning, early detecting anomalies that human eyes might easily miss, and leaving the final decision to the doctor or farmer. This article presents what is actually happening in 2025-2026, with specific examples from Vietnam and globally.

Part 1: AI in Medical Imaging Diagnostics

Why Medical Imaging is AI’s “Home Turf”

Medical imaging diagnostics — X-ray, CT, MRI, ultrasound, endoscopy — presents a task at which AI excels: the input is an image, and the output is a judgment of “presence or absence of abnormality, and its location.” Deep learning models are trained on hundreds of thousands of labeled images, allowing them to learn the subtle characteristics of tumors, lung nodules, fractures, or brain hemorrhages.

The maturity of this field is evident in regulatory figures. As of August 2025, the U.S. Food and Drug Administration (FDA) had approved 1,247 AI-integrated medical devices, with imaging diagnostics devices accounting for over 75% — specifically, X-ray/CT/MRI devices comprised approximately 723 of the total approved devices according to a systematic review. This makes it the most heavily “AI-ified” area of healthcare.

What AI Can Do, and How Well

For narrow and well-defined tasks, deep learning models have now achieved performance comparable to or surpassing that of human doctors:

• Fracture detection: Meta-analyses show pooled sensitivity and specificity both exceeding 90%, comparable to radiologists, making them sufficiently reliable as support tools.
• Lung cancer screening: A deep learning platform achieved an Area Under the Curve (AUC) of 93.6% for detecting invasive adenocarcinoma on chest images.
• Triage: AI automatically flags urgent cases like strokes for doctors to prioritize, shortening “door-to-treatment” times.

It is important to emphasize: the professional consensus in 2025 is that AI augments rather than replaces doctors. The best process is “human + AI”: AI quickly scans and suggests, while the doctor verifies and bears ultimate responsibility.

Real-world Implementation in Vietnam

Vietnam is not an exception to this trend. Some specific examples in 2025 include:

• Quang Ninh Provincial General Hospital (starting November 2025) integrated AI software to assist in reading chest X-rays. After receiving an image, AI automatically analyzes and detects abnormalities within approximately 3-10 seconds — significantly faster than traditional film reading processes.
• Bai Chay Hospital (Quang Ninh) invested in an AI image processing unit in its Diagnostic Imaging Department, supporting doctors in reading chest CT scans from early 2025.
• Southern Saigon International General Hospital continues to expand its application of AI in diagnostic imaging.
• From a technology enterprise perspective, VinBigData has developed and launched “pure Vietnamese” AI solutions for medical imaging diagnostics, tailored to the specific pathological characteristics and data of Vietnamese people.

Regarding policy, the Ministry of Health has issued a Digital Health Transformation Project, encouraging hospitals to adopt AI in diagnosis, treatment, and medical record management — creating a framework for technology to spread from central hospitals to provincial ones.

Significance for a Developing Country

For Vietnam, the greatest value of AI in healthcare lies in addressing the shortage of specialist doctors, particularly diagnostic imaging doctors at the grassroots level. A district hospital without an experienced radiologist can still use AI for preliminary screening and refer suspicious cases to higher-level hospitals — helping patients in remote areas access higher-quality diagnoses.

Part 2: AI in Precision Agriculture

What is Precision Agriculture?

“Precision agriculture” means using data for precise farming: watering where plants are thirsty, fertilizing where soil is deficient, spraying pesticides where pests or diseases are present — instead of uniformly applying across the entire field. AI acts as the brain, analyzing vast amounts of data from satellites, drones, soil sensors, and cameras to provide these recommendations.

The market is growing rapidly: according to IMARC Group, the AI in agriculture market in Vietnam reached US$8.72 million in 2024 and is projected to reach US$43.01 million by 2033, corresponding to a Compound Annual Growth Rate (CAGR) of approximately 19.4% per year during the 2025-2033 period.

Three Problems AI is Solving

1. Pest and disease diagnosis from images. Just as AI reads X-rays for humans, AI can “diagnose diseases” in plants through images of leaves, stems, or fruits captured by smartphones, cameras, or drones. In Vietnam, MiSmart has collaborated with the Plant Protection Department, utilizing AI-powered drones connected to the industry’s pest and disease database to monitor growth and detect early signs of pests and diseases.

2. Crop and yield forecasting. AI analyzes weather data, crop health, and soil conditions from the beginning of the season to predict end-of-season yields, helping farmers and businesses plan for harvesting, storage, and sales. In Thanh Hoa, some models applying AI forecasting systems for rice and corn have recorded an average yield increase of 10-15% compared to traditional farming, while also reducing losses due to pests and diseases.

3. Precision spraying and fertilization. AI drones designed by Vietnamese engineers are being used to spray pesticides according to maps, reducing chemical usage and labor. This represents a “tailored” solution approach for specific crop types and Vietnamese field terrains.

Global Examples: Benchmarking Possibilities

To visualize the potential, let’s look at global examples:

• John Deere (USA) with its See & Spray technology uses AI cameras to distinguish crops from weeds, spraying herbicides only on the weeds — reducing herbicide usage by up to 90%. This is the clearest demonstration of the “right place, right dose” principle.
• A wheat farm in Saskatchewan (Canada) uses satellite imagery and machine learning to predict disease outbreaks before visible symptoms appear, thereby reducing crop losses by approximately 30%.

These figures illustrate what Vietnamese agriculture is aiming for: AI transforming farming from “estimation-based” experience to data-driven decisions.

Common Denominators and Considerations

Both healthcare and agriculture demonstrate a common formula: AI excels at pattern recognition from images and data, but humans remain the ultimate decision-makers. Doctors are responsible for diagnosis; farmers decide on cultivation practices.

Some practical considerations for implementation:

• Local data quality: Models trained on foreign data may not align with Vietnamese pathologies, crop varieties, or soil conditions. Therefore, “pure Vietnamese” solutions like VinBigData or MiSmart hold particular value.
• Infrastructure and cost: Lower-tier hospitals and small farms require cost-effective, easy-to-use solutions that can function even with intermittent network connectivity.
• Legal responsibility: If AI provides incorrect suggestions, who is accountable? A clear legal framework is needed, especially in healthcare.

Overall, 2025-2026 marks the phase where AI in Vietnamese healthcare and agriculture transitions from laboratories to real hospitals and fields — slowly but surely.

References

• AI applied by hospitals in diagnostic imaging — VnEconomy
• Applying AI in healthcare, expanding opportunities for patient survival — Ministry of Science and Technology
• Applications of AI in medical imaging diagnostics in Vietnam — VinBigData
• AI in Radiology: 2025 Trends, FDA Approvals & Adoption — IntuitionLabs
• AI in Agriculture Market in Vietnam — IMARC Group
• Solving many “tailored” problems with AI drones designed and manufactured by Vietnamese — VietNamNet
• AI will help farmers forecast weather, pests, and product sales timing — Ho Chi Minh City Law Newspaper
• Ag Power John Deere: AI Products, 2025 Farming Trends — Farmonaut