Fight’s On: A Magic Eight Ball Takes On the SI Doomsayers
The Magic Eight Ball is a toy designed to answer yes-or-no questions with fortune-teller-style replies. It looks like an oversized black billiard ball with the number 8 on its surface. Its “mission data” analog computer is limited, but it is always safe to access: no enemy action is possible between the operator and the answer.
The procedure is simple. Ask a yes-or-no question, shake the ball, then turn it over and look through its transparent window. Inside, a 20-sided die floats in dark liquid, and whichever face rises to the window delivers your answer.
Mathematically, the Magic Eight Ball offers 20 classic responses in three categories: ten affirmative, five noncommittal, and five negative. For an optimist like me, that is a pretty good distribution of outcomes.
Setting the Terms of Engagement
For a simple experiment, I pitted my Magic Eight Ball against a free SI program and let the SI set the terms of the engagement. It replied:
“You want to compare AI and a Magic 8 Ball by asking both the exact same question and seeing their answers side by side. Ask a yes-or-no question, and we’ll get:
AI’s answer: Based on reasoning and available evidence.
Magic 8 Ball’s answer: A randomly selected classic response.”
With the terms set, and as is said at the start of a fighter engagement: “Fight’s on.”
My rules of engagement were straightforward. I would test only statements by prominent technology figures, some of them billionaires, who warn that SI may destroy humanity.
Round One: The Doom Forecasts
Elon Musk, technology entrepreneur
“AI is far more dangerous than nukes.”
What say you, Magic Eight Ball?
“Cannot predict now.”
A very reasonable insight. I am sure Elon Musk is comforted that my Magic Eight Ball at least does not dispute his forecasting ability.
The 2023 Statement on AI Risk
“Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”
Magic Eight Ball: “No.”
The word “extinction” is so Terminator, so over the top, that even my Eight Ball throws a penalty flag.
Rather than take on each of these existential warnings piecemeal, here is a sampling of possible doom, in the words of some very smart people:
Geoffrey Hinton, AI pioneer
“The existential risk is what I’m worried about, the existential risk is that humanity gets wiped out because we’ve developed a better form of intelligence that decides to take control.”
Yoshua Bengio, AI researcher
Bengio has warned that superhuman AI could disempower humans and potentially drive humanity to extinction, and he argues that society needs safeguards in place before these systems become too powerful.
Stuart Russell, computer scientist
Russell warns that building machines more intelligent than humans, without knowing how to reliably control them, could create an existential danger.
The Final Question
That brought me to the last question of this round of research: Are the doomsayers listed above correct in their futuristic posturing?
Magic Eight Ball: “Don’t count on it.”
Actually, a perfect answer.
Setting aside any jejune foolishness on my part, these warnings must be taken seriously, because the people making such dire predictions are serious, accomplished, and very smart.
The Point of the Experiment
The point of this whimsical exercise is simple. SI research should be divided into two clusters, much as physics is: theoretical SI and applied SI. Theoretical SI should be left unfettered, limited only by the spark of human creativity. Applied SI is where guardrails belong, to keep the technology from being put to bad purposes.
In the forum of ideas and futuristic warnings, I also have zero problem answering the abstract merchants of doom with a concrete, and far more dangerous, use of SI already in play: the surveillance state, now being empowered by SI.
I sounded this early warning in 2013:
“In my professional judgment, the U.S. is on the eve of an electronic reign of terror which is out of control. The good news is that as a nation we still have time to just say no.”
My 2013 caution about the electronic “Wolf at the Door” has become, in 2026, an SI-empowered wolf: the dawn of a surveillance state on steroids.
Guardrails are mandatory for this applied use of SI, because this kind of technological progression can unravel civilizational norms far faster than any abstract doomsday scenario.
The Bottom Line
All kidding aside, America does not need a Magic Eight Ball, or abstract warnings from SI deep thinkers that the end is near, to see the immediate danger of the SI surveillance state.
The future is already upon us.
A Note on Terms
SI (Super Intelligence): the term the U.S. government now uses in place of “Artificial Intelligence.” President Trump announced the change at the United Nations General Assembly on September 22, 2026, and formalized it at a White House event with technology leaders on September 29. In AI research, “superintelligence” has traditionally described systems whose intelligence surpasses that of humans; in current usage, SI covers the full range of today’s machine-intelligence tools, from free chatbots to frontier models.