Is there a special risk to GMOs?

"Mathematics, for all its unchallengeable power in framing theory, is tautological ... and it cannot inform us of the very special world in which we live." (Edward O. Wilson in "Consilience. The Unity of Knowledge.", p.68, Vintage,

Nassim Nicholas Taleb, the financial risk analyst and author of 'Antifragile' and 'The Black Swan' has recently published an account of why the precautionary principle (PP) from a pure mathematical perspective applies to genetically modified organisms, and that as a result of this analysis, GMOs should be banned absent evidence of their near-certain safety.

PP is used for risks that are thought to cause ruin or catastrophic failures to a system, unlike localized risks that can be managed using standard risk management (e.g. nuclear power plant melt-downs). The latter cannot spread through a naturally evolved complex system, ie. nature, and thus cannot threaten its survival. The question is, are GMOs systemic or localized risks. Taleb maintains the former although his own arguments about PP suggest otherwise. A key argument for PP with GMOs is the irreversibility of damage, like the extinction of a species, aka humans.

Inconsistencies in Taleb's argument relate to the following a priori assumptions:

- Evolution, including artificial selection (i.e. breeding) requires standard risk management, while top-down, human made GMOs require PP. Taleb gives no evidence why he deems genetic engineering top down (he does compare it to human made complex system, but this seems to me stretching the analogy between human made systems and life, i.e. a brain as a computer), and not bottom-up tinkering.

- Originally, one of his arguments against genetic engineering was that it is unnatural (which makes it dangerous and/or top down). This line of reasoning has disappeared in the current paper, and I believe correctly so. That genetic elements are crossing species and even kingdom (horizontal gene transfer), domain boundaries is well established biological trivia. So artificially inserting one or two bacterial genes into a plant, or a human gene into bacteria, is not an extraordinary event, except for the fact that any particular transfer chosen by genetic engineers would be unlikely to happen in nature, given chance and evolutionary constraints between the two chosen genomes by geneticists. The reason Taleb abandoned this argument is to avoid the circular problem in his attacks on GMO. You cannot proclaim, as Taleb does, that swapping genes as done by GMO does not happen in nature and at the same time claim that any such genetic manipulation is a system wide risk spreading throughout the living domain causing catastrophic failure, i.e., extinction.

- Note: originally the horizontal gene transfer across domains was deemed unnatural, now all the physicists supporting NNT's PP use horizontal gene transfer as the central argument against GMOs.

- And regarding the risk of GMOs spreading through all of nature, Taleb has no evidence how this would work (and simply claims that evidence is not needed and since no one knows the risk, it should be deemed beyond safe). You have to take him at his word, since Taleb offers not a single biological mechanism that would warrant his PP demand. He simply claims no evidence is needed, but here I would remind the reader of E. O. Wilson's take on the role of mathematics in science (see quote above).

I view his whole argument as an example of the paranoia principle, mostly based on rather weak logic. Just think about his comparison of the 2008 market crash and mass extinction. According to Taleb, the former is a fat tail (a black swan type risk to avoid), but the latter is not (local risk). To suggest that mass extinctions, which have seen a loss of more than 90% of species, as thin tailed risk, but market crashes and GMO as fat tails, is insane.

A technical aspect of Taleb’s argument has to do with his top down versus tinkering approaches. The latter is thin tailed, the former produces fat tails. Simply stating that GMOs are top down is not good enough. In fact, genetic engineering is tinkering with existing genomes, at a small scale, no less, since all successful GMOs include the changes of no more than a handful of genes. Compare this to a total gene content of tens of thousands in plants and animals. How this is top down boggles the mind.

In Taleb's latest defense against a well reasoned 'review' by Trevor Charles challenging his precautionary principle from a biological perspective, a lengthy comparison of GMO with 'natural' mutations (including those induced by artificial selection, aka breeding) paints a more modest argument against GMOs and includes lots of evidence, something the PP math simply is not based on. It is rather based on generic complex system, mostly human made, or top down engineered systems.