The recent surge of anxiety surrounding Artificial Intelligence, particularly regarding the potential for existential risks, often feels like a uniquely modern panic. However, a closer look at the history of computing reveals that warnings about machine supremacy are as old as the technology itself. As we edge closer to the reality of Artificial General Intelligence (AGI), it is crucial to analyze which historical predictions were overly dramatic and which have quietly become our reality.

70 Years of AI Warnings

The anxiety over AI dates back over seven decades. In 1951, computer scientist Alan Turing—often considered the godfather of AI—wrote a paper warning that once machines begin to think, it would not take long for them to “outstrip our feeble powers”. Turing famously proposed the “Turing Test” to determine if a computer could pass as human. He predicted this would happen around the year 2000. While his timeline was off by a couple of decades, today’s leading chatbots, such as those from OpenAI or Anthropic, have made distinguishing between human and machine text virtually impossible.

The warnings escalated over the decades:

• 1970: AI pioneer Marvin Minsky boldly claimed that AGI would be achieved within three to eight years, a prediction that spectacularly failed.
• 2000: California Congressman Brad Sherman publicly warned of a future where silicon engineering creates machines “considerably smarter than us who may or may not regard us as their appropriate peers or masters”.
• 2014: Renowned physicist Stephen Hawking sent shockwaves globally when he stated that the development of full AI could “spell the end of the human race,” warning that it would take off on its own and redesign itself at an ever-increasing rate. (A striking foreshadowing of the current focus on Recursive Self-Improvement).
• 2017: A landmark conference in California gathered over 100 leading voices in AI to establish ground rules, specifically warning against AI arms races and lethal autonomous weapons—issues we are grappling with directly today.

What Hasn’t Happened (Yet)

Despite decades of doomsday forecasts, some of the most common fears have not fully materialized. The prediction that AI would immediately take almost everyone’s jobs has not yet come to pass. We are not living in a world completely overrun by AI robots. Even driverless cars, while advancing rapidly, are still largely relegated to proving grounds and limited urban rollouts.

Furthermore, the ultimate fear—complete human extinction at the hands of a super-intelligent AI—remains a hypothetical, though highly debated, scenario among top researchers.

What Did Come True (And What Scares Experts Now)

While humanity has not been eradicated by machines, several critical predictions have materialized faster than expected. In 1997, IBM’s Deep Blue defeated chess grandmaster Garry Kasparov, achieving a major milestone a year ahead of projected timelines.

More alarmingly, the warnings from the 2017 conference regarding autonomous weapons are now a reality. Recent reports indicate that AI-controlled drones have autonomously killed targets in conflict zones without human input, proving that these systems can and do end lives in the physical world. Furthermore, AI systems are now capable of launching autonomous cyberattacks.

Another massive shift that was heavily predicted is the economic boom. AI spending is currently fueling an estimated 40% of US economic growth within the last year alone, driving unprecedented investment and computing power.

The immediate threat, according to tech analysts, isn’t necessarily a rogue super-intelligence waking up tomorrow. The real danger lies in highly capable AI systems falling into the hands of bad actors—such as hostile nation-states like North Korea or Russia. These tools could be used to seize control of electrical grids, manipulate water plants, or synthesize biological weapons before we ever reach true AGI.

As tech leaders like Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman continue to push the boundaries of what is possible, the core question is no longer if AI will change the world, but whether we have enough time to establish sufficient guardrails before it does.