The challenge with regulating technology is that laws always lag behind the technologies they regulate. One way to address this is to adopt a philosophy of incremental regulation that follows in lockstep with technological evolution. This requires a theoretical understanding of the evolutionary dynamics of technology.

This is a link-enhanced version of an article that first appeared in The Mint. You can read the original here. For the full archive of all the Ex Machina articles, please visit the website.

Modern technologies are modular, radically interoperable and interdependent. Which is why one of the foremost challenges in regulating them is dealing with complexity. I have long argued for principle-based regulations—in the belief that, if we express it at a high enough level of abstraction, our regulatory intent will be enforceable regardless of the complexity of the underlying technology or the direction in which it evolves.

But modern technologies are so inseparably integrated with each other and so internally complex that abstract principles alone may not be enough. We need to better understand how complexity works and exactly what needs to be done to address rapidly evolving technologies.

I recently came across a paper titled Assembly Theory by Leroy Cronin that offers a novel explanation for how complex systems evolve and grow. Even though it deals primarily with chemical systems, the ideas in this paper could not only apply to other complex systems but can also inform how we should regulate them.

Simply put, Assembly Theory describes objects by their formation histories—defining them in relation to their ‘assembly index’ (which goes by the steps required to build them from simpler components). One way to think about this is in relation to cooking. I know a simple recipe for biryani with just five ingredients that can be cooked in a pressure cooker. This is a simple dish with a low ‘assembly index.’ My grandmother, on the other hand, passed down a biryani recipe that had five pages of ingredients, all of which had to be prepared using eight complex techniques before they were assembled into the final dish. This is a complex preparation with a high ‘assembly index.’ Both recipes, if followed carefully, will deliver a biryani, but Assembly Theory posits that we need to evaluate the object based on its complexity rather than what the result looks like.

Regulation Assembly Index

This notion that the nature of an object has less to do with appearance and more to do with the core elements of a system and how they inter-operate with each other is a useful framework to assess technology systems in the context of their regulation. For this, we first need to break these systems down into their core components and assign them a regulatory assembly index (RAI) based on their complexity and functional interdependence.

Low RAI technologies—typically simple tools and standalone applications such as mobile applications or video games—require little more than light-touch regulation that focuses on basic safety and accountability. Medium RAI Systems— which are more deeply integrated multi-party platforms like e-commerce portals and ride-hailing systems—may need specific compliance protocols such as mandatory risk assessments and clearly defined obligations related to harms that could occur. High RAI Systems—typically critical infrastructure or population-scale systems such as Aadhaar or our digital payments infrastructure—call for more extensive oversight that could extend to pre-deployment testing, continuous monitoring and structured accountability measures.

That said, regulating technology based solely on its complexity takes us only part of the way. Given how rapidly technologies evolve, unless our approach to regulation takes this into account, it will come up short.

Compliance Selection

This is where ‘selection,’ the second aspect of Assembly Theory, becomes relevant. The paper’s authors argue that there is a process of natural selection by which circumstances other than complexity can determine why certain structures endure while others do not. Going back to the food analogy, when the two dishes described above are put on a restaurant menu, the less popular one will disappear from it over time. This notion that complex systems evolve in response to external factors provides yet another lens that can help us decide how to regulate technology systems.

The more complex a technology, the more likely that it will come under selective pressures. Complexity increases the number of components in a system and their functional interdependence. If we can design our regulatory frameworks to assess them at critical complexity thresholds, we might have better oversight over the direction of a system’s evolution.

Take social media apps, for example. At launch, they tend to be simple, offering a small set of features that users (and regulators) intuitively understand. But with time, new features get layered on till these applications become bloated and complex.

Incremental Regulation

If, instead of regulating the product or service as we currently do, we put in place a process that incrementally reviews these technologies at key milestones in their evolution, we will give regulators the ability to make small corrective adjustments as they evolve that will nudge these companies towards better data governance without stifling innovation.

The real benefit of Assembly Theory will be in the development of a governance framework that evolves in lock-step with technological complexity—one that continually refines its governance parameters to cater to increased complexity and selection metrics. If we can garner insights that help us balance the technological potential of new technologies with societal values, we will be able to not just regulate technology well but actively guide its evolution in a manner that would take us towards the desired beneficial outcomes.