Bill Gates and Elon Musk Are Wrong About AI: The Machines Are Not Taking Over the World, but They May Help Us Take Over Our Corner of the Universe
“If you think humans are going to be driven to extinction by something I can destroy with a bucket of water, then you’re an idiot. It’s time for AI companies to stop using fear to sell their products.” — Patrick Zarrelli
Artificial intelligence is going to change the world. It will eliminate jobs, create industries, accelerate scientific research, transform warfare, disrupt education and alter human civilization in ways nobody alive can accurately predict. But AI is not going to take over the planet, exterminate humanity and become Earth’s new dominant species. I would bet Elon Musk and Bill Gates $1 billion on it.
For years, some of the richest and most influential people in technology have described artificial intelligence using language borrowed from humanity’s darkest science fiction. Elon Musk famously warned that developing advanced AI was like “summoning the demon.” Geoffrey Hinton has contemplated significant probabilities of human extinction. Bill Gates has warned about humanity potentially losing control of sufficiently advanced systems. OpenAI CEO Sam Altman joined other prominent AI researchers and executives in signing a statement declaring that mitigating the risk of extinction from artificial intelligence should become a global priority comparable with preventing nuclear war and pandemics.
Those are extraordinary claims, yet they are increasingly treated as conventional wisdom despite the fact that nobody has demonstrated that technological intelligence naturally progresses toward machine consciousness, that machine consciousness inevitably develops self preservation, that self preservation inevitably produces hostility toward biological life or that intelligent civilizations ultimately abandon biology altogether. We have taken a stack of assumptions, placed another stack of assumptions on top of them and somehow arrived at human extinction.
Meanwhile, an entirely different technological revolution is accelerating underneath the AI frenzy: biotechnology. The long term future of advanced civilization may not belong to increasingly enormous warehouses filled with silicon processors. It may belong to stronger bodies, greater cognitive capacity, genetically engineered disease resistance, regenerative organs, dramatically extended lifespans, neural communication, biological computing and organisms whose technology has become indistinguishable from the organisms themselves.
Artificial intelligence may turn out to be one of the most important technologies humans ever create without becoming our successor. It may simply be the machine that teaches us how to rebuild ourselves.
Silicon Valley Has Been Predicting the AI Apocalypse for Years
Elon Musk delivered one of the most memorable AI warnings during an MIT symposium in 2014 when he declared, “With artificial intelligence we are summoning the demon.” Musk compared humanity’s relationship with advanced AI to stories in which someone believes he can control a supernatural force only to discover that he cannot. He has repeatedly returned to warnings that sufficiently powerful artificial intelligence could become dangerous to humanity.
Musk isn’t alone. In 2023, the Center for AI Safety released an extraordinarily concise declaration stating that mitigating the risk of extinction from AI should become a global priority alongside other civilization scale threats including nuclear war and pandemics. Among the signatories were some of the most influential people working in artificial intelligence, including OpenAI CEO Sam Altman, Google DeepMind CEO Demis Hassabis, Anthropic CEO Dario Amodei and computer scientist Geoffrey Hinton.
Hinton has subsequently discussed a significant possibility that AI could eventually contribute to human extinction. Think about the magnitude of what is actually being proposed. We aren’t talking about unemployment, deepfakes, autonomous weapons, cyberattacks or government surveillance. We are talking about human extinction.
Yet probabilities surrounding machine driven human extinction cannot be calculated from any meaningful historical dataset because no previous human civilization has documented the development of artificial superintelligence. These estimates are ultimately informed predictions about a technological condition humanity has never experienced. That doesn’t make the concerns worthless, but it does make treating speculative extinction probabilities as though they represent established scientific forecasts deeply questionable.
“Sometimes these AI systems can’t even tell you who the president is. When you correct them, they still get it wrong again tomorrow.” — Patrick Zarrelli
We Have Confused Intelligence With Consciousness
Artificial intelligence can already perform tasks that would have appeared almost supernatural twenty years ago. A modern frontier model can write software, analyze medical literature, translate languages, interpret photographs, search enormous datasets, generate video, solve mathematical problems and communicate across disciplines almost instantaneously. That represents extraordinary functional intelligence. It is not, by itself, evidence of consciousness or ambition.
Humans routinely project our own evolutionary psychology onto machines. We assume that if something becomes sufficiently intelligent, it will naturally want freedom. Then it will want survival, resources and power. Eventually, according to the familiar scenario, it will identify humans as competitors.
But those assumptions deserve scrutiny. Human beings possess fear because organisms that avoided danger survived long enough to reproduce. We compete for resources because biological organisms evolved in environments where resources were limited. We experience sexual desire because organisms without successful reproductive behavior disappeared from the gene pool. We defend territory, protect children, form tribes and fear death because natural selection spent hundreds of millions of years shaping those behaviors.
A computer did not evolve inside that environment. Intelligence does not automatically create hunger, jealousy, reproductive competition or hatred of human beings. Those characteristics would have to emerge from architecture, objectives, training, environmental interaction or deliberate human design.
The Terminator scenario therefore contains an enormous unstated assumption: once machines become intelligent enough, they somehow become psychologically human. That may eventually prove to be one of the strangest assumptions in the entire AI debate.
If You Had Never Seen Terminator, Would You Actually Be Afraid of AI?
There is a simple thought experiment missing from almost every conversation about artificial intelligence. Imagine you had never seen The Terminator. Imagine there was no Skynet, HAL 9000, Matrix or Ultron and Hollywood had never spent decades teaching us that once computers become intelligent, their natural next step is deciding human beings need to die.
Now open the artificial intelligence system you actually use today. Ask it to summarize a document, translate Spanish, analyze a spreadsheet, explain a scientific paper, help write computer code or compare competing arguments. Would your natural reaction really be that this technology is preparing to exterminate humanity?
Probably not.
The AI most people actually encounter is remarkably mundane compared with the monster living inside our cultural imagination. It helps people write emails, research subjects, generate images, translate languages, organize information, analyze data and automate repetitive work. It also makes mistakes, hallucinates information, misunderstands instructions and sometimes needs a human to correct something embarrassingly obvious. Then we close the browser tab.
That doesn’t make modern AI harmless. AI assisted fraud, deepfakes, cybercrime, surveillance, discriminatory automated decisions, labor displacement and weaponization are legitimate problems. But many of those dangers involve human beings using increasingly powerful technology against other human beings rather than an independently conscious machine deciding to attack humanity. That is an ancient problem wearing new clothes.
Facebook Never Became Conscious and Still Managed to Cause Real Problems
Social media provides a useful reality check because we don’t need to speculate about some hypothetical future version of it. Researchers, journalists, regulators, whistleblowers and technology companies’ own internal documents have spent years examining real world harms associated with social platforms, including misinformation, political polarization, extremist content, addictive design concerns and the effects of engagement driven systems on young users.
Facebook did not have to become conscious to become enormously consequential. It never woke up. It never developed a survival instinct. It never secretly decided it hated humanity. Humans designed an extraordinarily powerful information system, gave it commercial incentives centered around attention and engagement, deployed it across billions of people and then struggled with the consequences. That distinction should be at the center of the AI debate. The immediate danger is not necessarily that the machine becomes human. It is that humans use the machine to become more powerful.
A dictator equipped with sophisticated AI is concerning. A military deploying poorly controlled autonomous weapons is concerning. A criminal organization capable of automating fraud is concerning. A corporation capable of surveilling, profiling and psychologically targeting enormous populations is concerning. Those are human governance problems.
It is difficult to objectively quantify whether Facebook has caused “more harm” than artificial intelligence worldwide, but the comparison exposes the central weakness in the extinction debate. We have extensive evidence that existing digital platforms can produce serious societal consequences without becoming sentient, while the scenario in which AI independently develops consciousness, hostile intentions and the ability to eliminate humanity remains a prediction. The most dangerous part of an extraordinarily powerful tool may not be the tool suddenly developing evil intentions. It may be the human being holding it.
AI Has an Extraordinary Bird’s Eye View of Human Knowledge
There is also a tendency to confuse AI’s extraordinary access to information with omniscience. No human being can read every medical journal, court decision, scientific paper, historical archive, software repository, engineering manual and newspaper simultaneously. Artificial intelligence can increasingly operate across enormous portions of that accumulated information. That gives AI something resembling a bird’s eye view of human knowledge. It can identify relationships people miss because individual humans are trapped inside specialties, institutions, limited memories and limited lifespans. That capability is revolutionary, but the overwhelming informational foundation underneath these systems remains knowledge created, measured, written, photographed, calculated, recorded and discovered by human civilization.
Modern systems can absolutely derive new relationships from that information. They can produce solutions that were never explicitly written into their training material, perform sophisticated mathematical reasoning and increasingly use external tools to investigate problems. But that still does not make them unlimited general intelligences. Research continues to reveal failures when models encounter unfamiliar abstractions, increasing complexity and new environments. Apple’s researchers reported that frontier reasoning models could experience severe performance degradation beyond certain levels of problem complexity, even while acknowledging the substantial capabilities those systems demonstrate on other tasks.
The ARC AGI benchmarks tell an equally interesting story because that story keeps changing. AI systems struggled enormously with ARC AGI 2 when it was introduced in 2025. Frontier systems subsequently made extraordinary progress on the benchmark, demonstrating exactly how dangerous it is to declare any particular AI limitation permanent. Newer interactive reasoning environments, however, continue testing capabilities that cannot be reduced to impressive performance on previous benchmarks. That is what the evidence actually shows. Artificial intelligence is getting better incredibly quickly, but “getting better incredibly quickly” is not synonymous with “inevitably becomes a conscious autonomous god that kills us.”
The Next Great Technology May Not Be Mechanical
This is where the conventional AI narrative may be looking in completely the wrong direction. For most of human history, technology has meant external tools. A spear extends the arm. A telescope extends the eye. A car extends the legs. A telephone extends the voice. A computer extends the brain. Artificial intelligence dramatically extends our ability to process information.
Eventually, however, technological development may stop concentrating exclusively on improving the tool and start improving the human. Gene editing technologies are already giving scientists unprecedented control over genetic material. Regenerative medicine is exploring how damaged tissues and organs can be repaired. Brain computer interfaces are beginning to connect neural activity directly with external machines. Synthetic biology allows researchers to redesign biological processes, while researchers are experimenting with engineered tissues, organoids and biological computing.
The primitive version of the future is a human carrying an incredibly powerful computer. The advanced version may be a human who no longer needs to carry it.
Biology Is Already Technology Beyond Anything We Can Manufacture
Consider what a human body represents from an engineering perspective. A microscopic fertilized cell contains the instructions necessary to construct an organism containing tens of trillions of cells. That organism manufactures bones, muscles, blood vessels, eyes, skin, an immune system and a brain containing roughly 86 billion neurons. The resulting biological machine repairs itself, regulates its temperature, converts ordinary food into energy, identifies invading organisms, replaces damaged cells, learns and reproduces. The brain accomplishing the cognitive portion of this operates using remarkably little power compared with the industrial infrastructure supporting frontier artificial intelligence.
Now compare that with a modern AI data center. Data centers consume enormous quantities of electricity. Advanced processors require some of the most complicated manufacturing processes humans have ever developed. Semiconductor fabrication depends on ultrapure materials, extraordinary optical systems, massive industrial facilities and globally distributed supply chains. Cooling itself has become a major engineering challenge. Nature has been building self assembling, self repairing and extraordinarily energy efficient biological machines for billions of years. Why are we so certain an advanced civilization would throw all of that away?
AI Could Be the Tool That Unlocks Biological Super Technology
The more interesting technological trajectory may therefore be convergence. Artificial intelligence becomes extraordinarily good at understanding biology. It models proteins, analyzes genomes, predicts molecular interactions, searches enormous chemical spaces, helps scientists understand neural systems and accelerates drug development. Eventually it may help humans manipulate biological processes that are currently far too complicated for individual scientists to comprehend.
Biotechnology can then begin feeding improvements back into humans. Better disease resistance, regenerated organs, stronger bones, enhanced muscles, improved vision, expanded neural capacity, longer healthy lifespans and direct brain interfaces are all conceivable directions of research. Farther into the future could come forms of biological communication and computing that sound impossible from today’s technological perspective. At that point, asking whether humans or machines “won” becomes almost meaningless. The machine helped redesign the human.
The UAP Question Makes the Machine Future Even More Interesting
Then there is the elephant hovering over the entire conversation: unidentified anomalous phenomena. South Florida Media has spent years examining government disclosure records, military encounters, congressional testimony, declassified files and statements from senior officials concerning UAPs. The evidentiary record is substantially larger than the old caricature of someone photographing a strange light in the sky.
Military pilots have reported objects demonstrating unusual apparent performance characteristics. Naval personnel have described encounters involving multiple sensor systems. Members of Congress have publicly demanded greater disclosure. Former intelligence officer David Grusch testified under oath before Congress concerning alleged crash retrieval and reverse engineering programs. Retired Rear Admiral Timothy Gallaudet has publicly argued that the phenomenon is real and that important information concerning it has been concealed.
Readers do not have to accept every extraordinary UAP claim ever made to recognize that the disclosure record raises serious questions. One of those questions fits directly into the AI debate: if advanced nonhuman intelligence exists and has reached Earth, why do so many consequential allegations surrounding the phenomenon continue to involve biology?
Where Are All the Robot Aliens?
This is where the AI extinction argument becomes particularly fascinating. Suppose technological civilizations inevitably follow the trajectory envisioned by the most pessimistic AI theories. A civilization invents computers, those computers become intelligent, intelligent machines exceed biological intelligence, machines improve themselves, biological organisms become obsolete and eventually machine intelligence becomes the dominant form of advanced intelligence. If that progression is universal, sufficiently old civilizations should eventually become overwhelmingly machine civilizations.
So where are all the robot aliens?
The UAP disclosure world contains repeated allegations involving biological occupants, biological material and humanoid entities associated with extraordinarily advanced technology. David Grusch testified before Congress that “biologics” had allegedly been recovered in connection with some retrievals. When Representative Nancy Mace asked whether those biologics were human or nonhuman, Grusch said the assessment of people with direct knowledge was that they were nonhuman.
That testimony does not resolve the entire UAP question, but neither should sworn congressional testimony from a former intelligence official discussing information allegedly originating with people inside classified programs simply be waved away because the implications are extraordinary. If even part of the larger disclosure record ultimately proves accurate, humanity may be looking at something enormously important about technological evolution. Perhaps sufficiently advanced intelligence does not abandon biology. Perhaps it masters it.
Maybe the UFO Is Biological Technology Too
That possibility produces an even more radical interpretation. Humans separate machines from organisms because our technology is primitive enough that the distinction remains obvious. The pilot sits inside the airplane, the driver sits inside the car, the person holds the phone and the computer sits on the desk. But what happens after thousands, tens of thousands or millions of years of technological development?
The separation may disappear. A sufficiently advanced vehicle could theoretically incorporate biological materials capable of repairing themselves. A computational system could incorporate engineered neural tissue. Communication could become integrated directly into a nervous system. Bodies themselves could be genetically engineered. The distinction between pilot and craft, computer and brain, organism and machine could become increasingly meaningless.
What humans describe as a biological “alien” could theoretically be an engineered organism whose body, nervous system, communication abilities and lifespan have themselves become technologies. Even its apparent species could theoretically be designed. That remains a hypothesis. But the assumption that artificial intelligence becomes conscious, develops independent survival objectives, escapes human control and ultimately exterminates billions of people also contains layers of speculation. For some reason, Silicon Valley’s speculation gets international conferences and trillion dollar valuations while biological speculation is frequently dismissed as science fiction.
America’s AI Gods Still Need Taiwan to Manufacture Their Chips
There is another reason to inject some humility into the Silicon Valley narrative. The United States unquestionably leads important areas of frontier artificial intelligence, but America’s AI boom exists on top of an international technological system. NVIDIA designs extraordinary processors, but Taiwan Semiconductor Manufacturing Company manufactures many of the world’s most advanced chips. Advanced semiconductor production depends on specialized lithography technology and enormously complicated international supply chains. Japan remains critically important in semiconductor materials and manufacturing equipment. South Korea is central to advanced memory, while China remains indispensable across major portions of global electronics manufacturing and industrial supply chains.
The dependence became important enough that Washington committed billions of dollars through the CHIPS and Science Act to expand advanced semiconductor manufacturing inside the United States, including billions in direct federal support for TSMC’s Arizona expansion. The machine intelligence supposedly preparing to transcend humanity therefore still depends on an almost comically human arrangement of mines, ports, governments, engineers, factories, power plants, chemical companies, shipping networks and international alliances. Skynet has a supply chain.
“AI companies want you to believe their glorified typewriters can take over the world, but that’s just a distraction from them borrowing obscene amounts of money and gobbling up even more obscene amounts of resources.” — Patrick Zarrelli
Washington Is Helping Pay for the Revolution
The mythology of Silicon Valley also tends to portray technological progress as something brilliant entrepreneurs accomplish while government watches helplessly from the sidelines. Reality is considerably messier.
The United States has committed billions of dollars to rebuilding semiconductor manufacturing capacity domestically. Government agencies are enormous purchasers of technology, defense spending supports advanced research and public universities perform foundational science. Government funded research contributed to technologies underlying the modern internet, satellite navigation, computing and countless other systems subsequently commercialized by private companies.
Modern technology is the product of private enterprise, public investment and international supply chains working together. That makes it particularly ironic when billionaire technology executives emerge as political prophets explaining the future of civilization as though their corporations constructed modernity by themselves.
Fear Is Extremely Good for Business
There is also an uncomfortable economic incentive surrounding AI apocalypse rhetoric. Imagine telling investors that your company makes excellent software. Now imagine telling investors that your company is building the most powerful intelligence humanity has ever created and telling governments that the technology is so powerful it could potentially destroy civilization. Suddenly you aren’t selling software anymore. You are selling access to the future.
Your company becomes strategically indispensable. Governments need your executives in Washington. Militaries need your technology. Investors cannot afford to miss the revolution. Regulators need your expertise to understand the supposedly incomprehensible systems you created. Meanwhile, smaller competitors face the possibility that governments will decide artificial intelligence is simply too dangerous to allow everyone to develop freely.
None of this proves that Musk, Gates, Altman, Hinton or other prominent AI figures are lying about their concerns. Many researchers undoubtedly believe the risks are serious. But journalism is supposed to examine incentives alongside statements. When people connected to some of the most valuable technologies on Earth repeatedly tell humanity that those technologies may become unimaginably powerful, journalists, investors and governments should examine both the prediction and the economic ecosystem surrounding it.
The Real AI Dangers Don’t Require Killer Robots
None of this means artificial intelligence is harmless. Quite the opposite. AI poses immediate problems considerably more tangible than hypothetical robot extinction. Workers can lose careers. Governments can construct unprecedented surveillance systems. Corporations can automate discrimination. Authoritarian states can industrialize propaganda. Criminals can scale fraud. Deepfakes can make authentic evidence increasingly difficult to distinguish from fabrication. Autonomous weapons could shorten the distance between political decisions and lethal force.
Perhaps most importantly, a small number of corporations could accumulate extraordinary influence over the information infrastructure humans increasingly use to learn, work, communicate and understand reality. Those are serious dangers, and they are happening now. We do not need to invent mechanical demons to justify taking artificial intelligence seriously.
Humanity’s Future May Be More Human Than Silicon Valley Thinks
The biggest failure in the AI apocalypse narrative may ultimately be its lack of imagination. Silicon Valley imagines technological progress as a straight line in which computers become faster, then smarter, then smarter than humans, then autonomous, then dominant and finally make humanity obsolete. Technology rarely develops in straight lines. Artificial intelligence is simultaneously accelerating robotics, chemistry, physics, medicine, genetics, materials science and biotechnology. The breakthrough that changes civilization may therefore come from somewhere AI helps us reach rather than from AI itself.
Perhaps humans eventually live dramatically longer lives. Perhaps damaged organs regenerate. Perhaps inherited diseases disappear. Perhaps genetic engineering dramatically increases resistance to illness. Perhaps brain computer interfaces eventually become technologies incorporated directly into biology. Perhaps memory and cognition can be enhanced and communication becomes increasingly neurological. Perhaps our descendants would appear almost alien to us. And perhaps that is what genuinely advanced technology looks like: not humans replaced by machines, but humans transformed by knowledge.
Bill Gates and Elon Musk Are Wrong About the Endgame
Artificial intelligence deserves regulation, investment, serious scientific research and scrutiny from journalists who are not financially dependent on the companies building it. What it does not deserve is mythology. We have spent years listening to some of the world’s wealthiest technology figures tell humanity that their technological revolution may become so incredibly powerful that it could literally destroy civilization. Maybe they’re right, but there is remarkably little evidence establishing the entire chain of assumptions required to get from extraordinarily capable artificial intelligence operating inside data centers to the extermination of biological civilization.
There is, meanwhile, an entirely different technological frontier developing around genetics, synthetic biology, regenerative medicine, neuroscience and biological computing. Hanging over that debate is perhaps the strangest clue of all: if the emerging UAP disclosure record ultimately demonstrates that humanity is interacting with an intelligence vastly more technologically mature than our own, we should pay extremely close attention to what that intelligence actually became. Maybe advanced civilization doesn’t discard biology. Maybe it perfects it. Maybe AI doesn’t replace us. Maybe AI helps us become something better. Elon Musk and Bill Gates can keep betting on the machines.
I’ll put my billion dollars on biology.
Sources
• South Florida Media — UFO Disclosure File Viewer
• South Florida Media — UFO and UAP News Archive
• U.S. House Committee on Oversight — 2023 UAP Congressional Hearing
• David Grusch — Written Testimony to Congress on UAP Programs
• U.S. House Committee on Oversight — 2024 UAP Hearing Featuring Rear Admiral Timothy Gallaudet
• Rear Admiral Timothy Gallaudet — Congressional Testimony on UAP Encounters and Secrecy
• South Florida Media — Rear Admiral Timothy Gallaudet Says U.S. Has Been Hiding UFO Data for Decades
• Center for AI Safety — Statement on AI Extinction Risk
• Bill Gates — The Choices We Make About AI Now Are Critical
• ARC Prize — Current ARC-AGI Benchmark Results
• U.S. Department of Commerce — $6.6 Billion TSMC Arizona CHIPS Award
