Artificial Intelligence and Real Information

Example 3: Logic Checking

Can AI be used to increase the amount of false and misleading information? Of course. Will it be used that way? Already is. More interesting and more important, can AI be used to improve the quality of information? Absolutely, so long as we know how to use it that way.

There was an article in The New York Times headlined Surprising New Research Links Infant Mortality to Crashing Bat Populations. “Bats eat insects. When a fatal disease hit bats, farmers used more pesticides to protect crops. And that, according to a new study, led to an increase in infant mortality.”

Whether and to what extent various agricultural chemicals affect human health is an important and debated question. The answer is obviously of much more than academic interest, but most experiments to rigorously test these effects would be ethically unacceptable. The Times article says that the article from Science it describes drew its conclusions from a rigorous study of an unintentional experiment staged by nature, the decline in bat populations due to white-nose syndrome, a fungus. Less bats mean more bugs. More bugs mean farmers use more insecticides. In the places where the bat die-off is significant, there is also an increase in infant mortality due to causes other than accidents or murders. According to The Times “The new study tested various alternatives to see if something else could have driven the increase: Unemployment or drug overdoses, for example. Nothing else was found to cause it.”

We all know better than to put much faith in a reporter’s interpretation of a scientific paper, but the article does link to the actual study: “The economic impacts of ecosystem disruptions: Costs from substituting biological pest control”. Because I would like to know what the health effects of various insecticides are, I read the original article. Although I didn’t find any errors in the math nor the general methodology (both of which are well-presented in the article), the article doesn’t present any direct cause and effect relationship between the use of insecticides and infant mortality but rather relies on the assertion that there doesn’t seem to be any other credible explanation. “I demonstrate how declines in insect-eating bat population levels induce farmers to substitute with insecticides, consequently [emphasis mine] resulting in a negative health shock to infant mortality,” the author asserts.

But are there other credible explanations which should be explored? I asked ChatGPT: “https://www.science.org/doi/10.1126/science.adg0344 the linked-to article hypothesizes that increased insecticide use due to an increase in insects due to a decline in the bat population because of white-nose disease is the cause of increased infant mortality in the areas where white-nose disease is prevalent. Accepting as fact that there are increased insect populations where white-nose disease is prevalent and that there is also an increase in infant mortality in those areas, what could be causes of the increase in infant mortality other than or in addition to the insecticide use.”

ChatGPT’s conclusion (full answer below): “Increased insecticide use might be a contributing factor to the rise in infant mortality in areas affected by white-nose syndrome in bats, but it is likely not the only cause. Other factors, such as disease transmission from insects, environmental degradation, crop damage, malnutrition, and healthcare access, could all play a role in the observed increase in infant mortality. These factors may interact in complex ways, making it necessary to consider a broad range of potential causes when examining changes in infant health outcomes.”

I then asked: “does the referenced article consider that these other factors may be leading to increased infant mortality?”

ChatGPT: “The article in question, which explores the connection between bat population decline due to white-nose syndrome (WNS), increased insecticide use, and a rise in infant mortality, primarily focuses on the hypothesis that increased pesticide usage is the key factor driving higher infant mortality in affected areas. The researchers argue that the collapse of bat populations, which naturally control insect populations, led to a surge in insecticide use by farmers. This increased use of chemicals is linked to a significant rise in infant deaths, particularly in rural areas where bats previously played a crucial role in pest management.

“However, the study does not deeply explore alternative factors that might also contribute to the increased infant mortality, such as the potential for higher disease transmission from the increased insect populations or environmental factors like water contamination. The authors focus mainly on the economic and health costs of increased insecticide usage, but the broader ecological disruptions caused by declining bat populations, including impacts on food security or increased exposure to vector-borne diseases, are not central in the discussion​( ScienceDaily, Phys.org).

“Thus, while the study strongly emphasizes the pesticide-infant mortality link, other factors like food security, disease transmission, and environmental degradation could also contribute to increased infant mortality in these areas and merit further investigation.”

Why does this matter?

If we uncritically accept the results of the study, we might reasonably lower the use of insecticides even at the cost of decreased food production. But what if the insects and not the insecticides are mainly to blame for the increase in infant mortality. Less insecticides then might lead to even more infant mortality, hardly a desirable outcome. Back in plague days some experts convinced people in some towns that cats caused the plague. They killed the cats. The rats, which carry the fleas which transmit the disease, multiplied. Whoops.

None of this proves that insecticides are not the cause of excess deaths. The article makes a convincing case that there are many bad consequences from the bat die-off including health effects. The statistics do show that the health effects of insecticides need further study; insecticides are not convicted, but they’re not exonerated either.

I know next to nothing about the health effects of insecticides, a highly debated and important subject. I do know that AI can help us lay people evaluate evidence on which we have to make decisions. AI can help us know what we know... and what we don’t know. AI, properly used (and, in this case, easily), will help us make better decisions. Education for us adults and out children must include how to use AI to improve the quality of information.

 See also:

Example #2: Fact or Fiction, the GPT.

Example #1: How To Get Useful Answers with ChatGPT Plus

Full text of AI “thoughts” on possible cause of infant mortality related to bat die-off (tl;dr)

Q: https://www.science.org/doi/10.1126/science.adg0344the linked-to article hypothesizes that increased insecticide use due to an increase in insects due to a decline in the bat population because of white-nose disease is the cause of increased infant mortality in the areas where white-nose disease is prevalent. Accepting as fact that there are increased insect populations where white-nose disease is prevalent and that there is also an increase in infant mortality in those areas, what could be causes of the increase in infant mortality other than or in addition to the insecticide use.

 A:

1. If we accept that both increased insect populations and increased infant mortality are observed in areas affected by white-nose syndrome (WNS) in bats, there are several potential causes of the rise in infant mortality beyond or in addition to increased insecticide use. These potential factors could be environmental, ecological, or public health-related, and might interact in complex ways with the decline in bat populations and the subsequent insect surges.
2. Increased Disease Transmission from Insects:Vector-borne diseases: An increase in insect populations, particularly mosquitoes, could lead to a rise in diseases transmitted by insects, such as malaria, dengue fever, or West Nile virus. Insect-borne diseases can affect infants directly, contributing to higher mortality rates, especially in areas with poor healthcare infrastructure.

• Pathogen amplification: Insects can carry and transmit various pathogens, including bacteria, viruses, and parasites. With increased insect populations, the transmission rates of these pathogens might rise, affecting vulnerable populations like infants. T

2. Nutritional Impacts:

• Crop damage due to increased insects: With the decline in insect-eating bats, there could be more crop damage from insects like locusts or beetles. This might result in reduced agricultural yields, leading to food shortages or reduced food quality, which in turn could affect infant nutrition, leading to malnutrition or food insecurity.

• Increased reliance on processed foods: If local crops are more prone to insect damage, communities may turn to cheaper, processed foods with lower nutritional value, further impacting infant health and development. 

3. Environmental Degradation:

• Water contamination: Increased insecticide use or environmental runoff from farmlands due to higher insect populations could contaminate local water sources. Contaminated water can lead to waterborne diseases such as diarrhea, cholera, and other infections that are particularly dangerous for infants.
• Air pollution: Spraying insecticides could also increase air pollution, particularly if chemicals are aerosolized. Poor air quality may exacerbate respiratory issues in infants, who are more vulnerable to air pollutants. These are both due to pesticides

4. Immune System Impacts:

• Immune stress from environmental toxins: Insecticides, even in low concentrations, may cause stress to the immune systems of infants, either directly or through the mother's exposure during pregnancy. This immune stress could make infants more susceptible to infections or other health complications.
• Endocrine disruption: Many pesticides are known to be endocrine disruptors, which can affect infant development and overall health, potentially contributing to conditions such as developmental delays or birth defects. These are all due to pesticides and the evidence is rapidly accumulating Due to pesticides

5. Socioeconomic Factors:

• Increased healthcare burden: Areas dealing with both ecological changes (increased insects) and increased pesticide use may also face economic strain, reducing the capacity of public health systems to adequately care for infants. This can lead to poorer healthcare access, delayed treatment, or inadequate maternal and infant care.
• Economic disruption: Increases in insecticide use and crop damage could lead to economic downturns in agricultural areas, reducing household incomes. This can affect access to healthcare, clean water, and nutritious food, all of which are crucial for infant survival. There is little doubt that the cause is complicated and multifactorial

6. Displacement of Predator Species:

• Imbalance in the ecosystem: The loss of bats could lead to broader ecological imbalances. Other insect predators, such as birds, frogs, or small mammals, may also be affected by declining insect prey or habitat changes, leading to more drastic fluctuations in the insect population. This broader ecological disruption could have indirect consequences on food security, disease ecology, or water contamination, all of which may affect infant mortality rates.
  

7. Secondary Effects of Bat Population Decline:

• Altered ecological services: Bats play a critical role in pollination and seed dispersal in some ecosystems. Their decline could affect local flora and agricultural ecosystems, possibly reducing biodiversity or changing the availability of certain crops that support community nutrition.
• Impact on zoonotic diseases: The decline of bat populations might disrupt ecological checks on certain pathogens. Bats play a role in managing insect populations that may carry zoonotic diseases, and their absence could lead to increased transmission of diseases that indirectly affect infant health.

8. Climate Change or Environmental Stressors:

• Amplifying effects of climate change: If climate change is contributing to the increase in insects, either through warmer temperatures or altered precipitation patterns, it could create a feedback loop that worsens infant mortality. Climate change may also exacerbate other stressors, such as food security or disease transmission.

Conclusion:

Increased insecticide use might be a contributing factor to the rise in infant mortality in areas affected by white-nose syndrome in bats, but it is likely not the only cause. Other factors, such as disease transmission from insects, environmental degradation, crop damage, malnutrition, and healthcare access, could all play a role in the observed increase in infant mortality. These factors may interact in complex ways, making it necessary to consider a broad range of potential causes when examining changes in infant health outcomes.