Experts, philosophers, academic institutions and governments are sounding the alarm bells in an effort to engage humanity in a conversation about the prospective emergence of superintelligent entities. They have posited, with varying degrees of certainty, what might happen in a world with increasingly intelligent machines. Wildly differing predictions have captured the human imagination — from certain, existential doom to infinite utopia.
Some claim these machines will behave parasitically, like super-powered viruses, relentlessly optimizing their reproduction and stamping out all other life on Earth. Others believe they may help us enter a post-scarcity era in which food, income, shelter and energy are universally provided for an increasingly healthful and long-lived population.
But much context is missing from the conversation. If we’re anticipating historic shifts in the way our societies run, shouldn’t we try to understand how we responded to previous major transformations to our societies, such as those brought on by agriculture and industrialization? Shouldn’t we try to anticipate the conditions that will facilitate the emergence of superintelligent entities?
Surprisingly, there’s been no public discussion that includes the appropriate context. But this is arguably as important to society as climate change, and quite possibly happening on a faster timeline.
Through my research on the subject, I’ve identified several thought experiments and hypotheses I believe can advance human discourse on the subject. But we need to steer the conversations beyond the sci-fi provoked questions to delve into more substantive issues. For example:
1. Inflammatory ideas are running amok. Much of humanity’s confusion about the arrival of machines with superintelligence arises from our difficulties in understanding their most relevant traits and identifying them in a way that facilitates productive discourse. It’s not uncommon for our problem-solving discussions to get knocked off track by half-baked ideas and ill-fitting terms. And that’s exactly what is happening. Many non-experts are inflaming public opinion by employing nonsensical arguments. Many experts are confusing public opinion about the emergence of superintelligent machines without providing appropriate context.
A common tactic is to show a robot with similar anatomy to humans and to insinuate that evil would emerge from it, as it has in evil humans of the past. Of course, this does nothing to introduce the expansive capacities of these entities in a way that advances public discourse.
2. The mainstream utilizes confusing nomenclature. Smarter, more philosophical experts have introduced the idea of “superintelligence,” which they define as “any intellect that greatly exceeds the cognitive performance of humans in virtually all domains of interest.” While this term is warranted, as it indicates superintelligent entities will compensate for one of humanity’s two cognitive impediments — our data-analysis limitations, it doesn’t indicate that they’ll also compensate for our other cognitive impediment — our illusory self, which keeps us disconnected from humanity’s real needs.
With each narrowly focused Artificial Intelligence that merges into the codebase of an emergent superintelligent entity — whether a mental-health diagnostic algorithm, a financial trend-seeking tool, an electric-grid-management platform or an intelligent personal assistant — these new entities will be aware of an increasingly wide and diverse domain in ways any single human, or even a large collaborative of humans, never could. This is why I’ve introduced the term “super-awareness.”
Super-awareness indicates these machines will have real-time awareness of seemingly infinite data and knowledge, while superintelligence indicates they will know how and when to use it to produce new knowledge — examining, comparing and calculating the probabilities of millions of different scenarios at the same time, across all human disciplines and industries. I believe combining the term “super-awareness” with superintelligence, Super-Aware/Intelligent Machines, is a valuable way of identifying these machines. It speaks to the far-reaching nature of their capacities, as compared to the siloed, limited intelligence of humans and our institutions.
When these entities emerge, it seems to me that they’ll be, above all else, Super-Aware/Intelligent Machines. By moving beyond apocalyptic prophecies and restrictive labels, and striving to better understand the potential implications of Super-Aware/Intelligent Machines on our world, we can anticipate the implications in a more productive way.