Summary of Key Points
This news article focuses on the topic of “how to use data to manage planetary health.” The central argument is that the current bottleneck in research on planetary health is not a lack of data, but rather the fact that data is “too dirty” (unorganized and incompatible) to be effectively utilized. Professor Schieder has initiated the “Planetary Health Axis System (PHAS),” which utilizes AI technology to integrate multi-dimensional unstructured data. This system employs a “validation loop” to clean the data and identify complex patterns. However, it emphasizes that decision-making power must remain in the hands of humans. Additionally, challenges such as the communication gap between traditional scientists and AI experts, as well as the resource constraints in developing countries, need to be addressed in order to effectively tackle complex issues related to climate, food, and health.
Detailed Analysis
1. The Problem is Not a Lack of Data, but “Dirty Data”
Many people believe that the lack of data hinders research on the relationships between climate, food, and diseases. Schieder argues that the real issue lies in the poor quality of the data—satellite temperature data, hospital records, and crop monitoring information are often in disorganized formats (text, images, sensor readings) that are incompatible with traditional scientific methods, which rely on structured datasets. Even worse, 80%-90% of the world’s data is unstructured, making it essentially unusable despite being abundant.
2. The PHAS System: Using AI to Clean “Dirty Data” and Establish a “Validation Loop”
The PHAS system is designed to address this problem:
- Step 1: Integrating Multi-Dimensional Data with AI: It combines data on climate, ecology, food, public health, and more (even unstructured data), covering over 48,000 variables that encompass various aspects of planetary health.
- Step 2: Cleaning Data through a Validation Loop: After collecting the data, the system runs models repeatedly, comparing the results with authoritative academic papers and expert judgments to eliminate irrelevant information and optimize the data structure. After several iterations, the data becomes cleaner, and the models more accurate.
3. The Power of AI to Uncover Hidden Patterns
While humans can consider only a few factors at a time, AI can process thousands of variables and identify complex relationships. For example, extreme heat can lead to reduced crop yields, which in turn may trigger food safety crises, with underlying factors including soil quality, policies, and trade. AI can uncover these hidden patterns in large datasets and even reveal new connections that were previously unnoticed.
4. Decision-Making Should Not Be Entrusted to AI: Tools Are Supplementary, Humans Are the Key
Some worry about who will be responsible if the PHAS system makes a mistake. Schieder clarifies that AI is a tool, not a replacement for human decision-making. The PHAS system is interactive, allowing policymakers to simulate different scenarios (e.g., whether a policy from Region A would work in Region B). Ultimately, humans must make the final decisions, as financial, environmental, and social conditions vary from place to place, and responsibility also lies with humans.
5. Two Additional Challenges to Overcome
- Communication Barriers Between Experts: Traditional scientists and AI experts often operate in silos (attending different conferences and publishing in separate journals), and some journals do not have dedicated review processes for AI research. Schieder suggests that mainstream journals should publish special issues to facilitate communication between the two groups.
- Resource Constraints in Developing Countries: Southern countries lack the resources and technology to adopt either the European Union’s “regulatory-first” or the United States’ “market-first” approaches. They need to build a solid data foundation first and use advanced tools to address their specific challenges.
The core message of this news is that while AI can help solve the problem of dirty data, humans must ultimately make the decisions. Additionally, breaking down interdisciplinary barriers and assisting developing countries are essential for effectively using data to manage planetary health.