The Man Who Created Artificial General Intelligence
▼ Summary
– The term “artificial intelligence” was coined by John McCarthy at a 1956 Dartmouth College meeting where academics discussed making machines think like humans.
– Artificial general intelligence (AGI) refers to computers matching or surpassing human intelligence and is a major driver of current tech investments and geopolitical concerns.
– Mark Gubrud first introduced the term “artificial general intelligence” in a 1997 paper on nanotechnology and international security.
– Gubrud defined AGI as systems rivaling the human brain in complexity and speed that can acquire, manipulate, and reason with general knowledge.
– His original definition, minus a clause about industrial or military use, closely aligns with how AGI is understood today.
The quest for Artificial General Intelligence (AGI) represents one of the most ambitious technological pursuits of our time, aiming to create machines with cognitive abilities that equal or exceed human intelligence. While the broader concept of artificial intelligence was famously named at the 1956 Dartmouth Conference, the specific term “artificial general intelligence” and its modern definition have a much more obscure origin story.
Today, AGI dominates technology headlines, fueling massive corporate investments from giants like Microsoft, Meta, and Google. The intense competition to achieve this milestone first has even been framed as a matter of national security, with significant geopolitical implications. This drive has also propelled companies like Nvidia to unprecedented valuations. Despite the term’s current prominence, the individual who originally coined “AGI” and established its core meaning remains largely unknown to the public.
That person is Mark Gubrud. Back in 1997, his primary focus was not on AI, but on the potential dangers of nanotechnology. As a graduate student at the University of Maryland, he worked in less-than-ideal conditions, describing his workspace as a “sub-sub basement” with a loudly cycling sump pump right behind his desk. There, he immersed himself in research about emerging technologies and their risks, heavily influenced by the work of Eric Drexler, a leading voice in nanotechnology.
Gubrud’s concerns centered on how such powerful technologies could be weaponized, creating threats even more catastrophic than nuclear warfare. He actively participated in scientific conferences to voice these warnings. At the Fifth Foresight Conference on Molecular Nanotechnology, he presented a pivotal paper titled “Nanotechnology and International Security.” In it, he argued that breakthrough technologies would fundamentally reshape international conflict and urged humanity to move beyond its “warrior tradition.”
Within this paper, while discussing the perils of advanced technologies, he introduced a then-novel phrase: “artificial general intelligence.” Historical records suggest this was its first known usage. He proceeded to define it with remarkable foresight, writing that it referred to “AI systems that rival or surpass the human brain in complexity and speed, that can acquire, manipulate and reason with general knowledge, and that are usable in essentially any phase of industrial or military operations where a human intelligence would otherwise be needed.”
If you remove the final clause concerning military and industrial application, his definition aligns almost perfectly with how AGI is commonly understood today. Gubrud’s prescient definition established the foundational goal for the field: creating an AI that can understand, learn, and apply its intelligence to solve any problem a human can. His early conceptual work laid the groundwork for a technological race that now captivates the world’s largest tech corporations and governments.
(Source: Wired)
