Dual uses of synthetic biology – time to take a step back?

As we’ve seen with vaccinations and other areas in global health, biomedical research, when carried out properly, can save millions of lives. However, the incredible speed and power of scientific and technological development can also be turned against us to do great harm if left unchecked.??

One area that deserves increasingly thoughtful consideration is the field of synthetic biology, especially when viewed alongside the growing power of artificial intelligence and machine learning.?

Gordon Moore co-founded a company in the 1960’s called Fairchild Semiconductor and later founded Intel.??In 1975, he forecasted a doubling every two years in the number of components per integrated circuit.??Over time, this became known as “Moore’s Law ”.??We can see the impact of Moore’s law in the many advancements in things like micro processing speeds, better and better digital cameras and other incredible breakthroughs.??There’s a reason why advancement seems to be getting faster and faster!

Now take the field of synthetic biology, which has recently become widely known by the sequencing of the human genome and the development of CRISPR.???Synthetic biology bridges engineering and biology.??For example,?one use case is our ability to regulate??cellular networks by constructing artificial synthetic targets, which brings with it incredible opportunities to treat diseases with great precision .?

Just last week, Google’s Artificial Intelligence division, DeepMind,?announced ?that it was able to – for the first time – predict and understand the structure of almost every known protein on earth. Proteins are the building blocks of life.?

As the?Financial Times reported last week , their database now contains over 200 million protein structures, “covering almost every organism on Earth that has had its genome sequenced – from the malarial parasite to the honeybee.”??All of these structures are now available via a public database hosted by the European Molecular Biology Laboratory (EMBL-EBI).??Dr. Janet Thornton, of the EMBL-EBI noted to the FT, “Almost every drug that has come to market over the past two years has been in part designed through knowledge of protein structures”.

With an understanding of protein structures, combined with the ability to “hit print” on a gene synthesizer, the ability to “create life” is fast becoming reality.??

Working with a social change organization I am part of called? Helena , my team and I began to wonder, might there be uses of this technology that could be pointed in the wrong direction and potentially cause irreparable harm?

Transparency and “open sourcing” have been key tenets of science (e.g. scientific publishing) and technology for decades.??But, might the publication of gene sequences - however well-meaning - in a world of decentralized gene synthesizers, result in catastrophic risk??

We then thought back to Gordon Moore, and considered whether the application of artificial intelligence to synthetic biology might compound two exponential risk curves.???

The knowledge to fold all proteins, for instance, is exceptionally powerful - and for purposes that are not all in the service of good. Even if never weaponized, these technologies could portend catastrophic risk simply via what are likely to be inevitable accidents.??Are we reaching a point where what was once kept locked away in Biosafety Level 4 laboratories (like smallpox) might now be able to be reverse engineered through easily accessible information, printed, and then accidentally leaked?

For instance, according to a?recent publication from the United Nations Interregional Crime and Justice Institute (UN-ICRI) , “malicious actors might use protein-folding prediction algorithms for purposes unintended in their original development. Understanding how genetic variation affects protein structure may enable researchers to discover disease-causing mutations using one individual’s genome sequence. While this ability might raise serious concerns for individuals who carry rare and/or severe genetic diseases, the threat would become even more significant if malicious actors could predict how genetic variations impact the severity of diseases and identify individuals susceptible to specific pathogens. These threats open the door to a dual-use technology dilemma and raise the question of how we can prevent malicious actors from weaponizing protein-folding prediction while still fostering scientific advancement.”

There are efforts being considered, for instance, to collect the genomes of thousands of undiscovered viruses, and publish them online so that the world’s researchers might be able to find cures for them in advance.??The motivation to do so is likely of pure intent and laudable.

However, as Jonas Sandbrink, Janvi Ahuja and colleagues from Oxford and other leading universities argued in a?recent paper , “Once a pandemic-capable pathogen has been identified, its genome features high dual-use potential: it may inform biosurveillance while also constituting a blueprint to cause widespread harm.”??Once we identify and alert the public to a “table of contents” of viruses, are we creating an?information hazard , perhaps inadvertently contributing to a potentially catastrophic event that might infect billions of people?

The magnitude of this risk is multiplied by the rate of its advancement is many orders of magnitude more dangerous than all previous global health issues combined.??The world is likely deeply unprepared for this, and the current damaged state of multilateral collaboration and unity across countries will only add fuel to the fire.??

I believe it is time that we give urgent but thoughtful consideration to the rate of the technological progress that is doing so much good for our world, and deeply consider the downside risks of decentralizing the power to create potentially catastrophic synthetic biology.??

A clear process is required that might bring the best minds from the multilateral system - such as the United Nations and its specialized agency for health, the World Health Organization - governments, the private sector, and cross-disciplinary academics - together to determine a framework for managing catastrophic biological risks.??This framework might include when open-sourcing efforts and decentralized access to certain tools may potentially risk more harm than good.??I would encourage us to keep all options on the table.

With the COVID-19 pandemic still fresh on our minds, and with millions that will likely continue to die from it, might this be a moment to “press pause” and look deeply at technologies that might enable the past two terrible years of suffering to resemble a dress rehearsal.??

Special thanks to Suprotik Basu and Daniel Schmachtenberger for their contributions to this article.