This article first appeared in the World Economic Forum's Global Risks 2018 Report.
As the world becomes more complex and interconnected, easily managed incremental change is giving way to the instability of feedback loops, threshold effects and cascading disruptions. Sudden and dramatic breakdowns— future shocks—become more likely. In the Global Risks 2018 Report, we present 10 such potential future shocks. Some are more speculative than others; some extrapolate from risks that have already begun to crystallize. These are not predictions. They are food for thought and action—what are the possible future shocks that could fundamentally disrupt or destabilize your world, and what can you do to prevent them? Below, is one potential future shock regarding the evolution of artificial intelligence.
What if the adverse impact of artificial intelligence (AI) involves not a superintelligence that takes control from humans but “AI weeds”—low-level algorithms that slowly choke off the internet?
Algorithms are already proliferating. As they increase in sophistication—as we become more reliant on code that writes code, for example—explosive growth becomes more likely. A divergence could open between the code we have created and our capacity to track and control it. The tragedy of the commons means we often let chronic problems with dispersed responsibilities fester. Think of plastic in the ocean. A trend towards reduced internet efficiency would undermine service delivery in countless businesses. It could hobble the Internet of Things. It would frustrate users. If the problem became significant enough, it could prompt some governments to wall off parts of the internet. If malicious actors found ways to proliferate or weaponize the AI weeds, they could do extensive damage.
As the global demands placed on the internet increase in scale and sophistication, digital hygiene is likely to become a more pressing concern for endusers. The development of overarching norms, regulations and governance structures for AI will be crucial: without a robust and enforceable regulatory framework, there is a risk that humans will in effect be crowded out from the internet by the proliferation of AI.
