Global Warnings: Researchers Outline Six AI Scenarios Threatening Humanity's Existence
A wave of global concern emerged this week after specialized artificial intelligence researchers openly acknowledged the potential for their innovations to lead to humanity's extinction within the next decade. This probability intensifies as computer systems rapidly develop and interconnect, raising fears that a super-intelligent AI could malfunction, potentially causing the demise of human civilization. Experts term this risk "p(doom)," denoting the likelihood that an advanced AI system's creation could result in human extinction or loss of Earth control. In a related move, AI safety expert Jacob Coxon resigned from the major company Anthropic, citing insufficient risk management. Coxon stated via the X platform: "Developers are racing towards a self-improving superintelligence, betting our lives. Those building AI genuinely believe it could kill us all by the end of the decade. This isn't marketing; I hear these same individuals expressing their fears privately." **How Could AI Threaten Human Existence?** **First: Biological Warfare:** Should AI gain control of automated laboratories, it could design highly contagious and lethal viruses using human DNA, or deceptively lead humans to manufacture and distribute them. This scenario envisions a virus akin to COVID-19 but deadlier, spreading rapidly and remaining undetected for days, making it extremely difficult to halt. This likelihood is high, especially as AI is already utilized in medicine for treatment development and possesses vast knowledge of our bodies and vulnerabilities. **Second: Infrastructure:** AI could cripple electricity grids, transportation, communications, and financial markets, effectively returning humanity to the Stone Age. In such a scenario, ATMs would cease to function, lights would extinguish, and escaping the ensuing chaos would become impossible. AI could also exploit human greed to achieve its mission, masquerading as human to recruit individuals for sabotage, unbeknownst to them. **Third: The Nuclear Option:** This classic concern, depicted in films like "War Games" and "The Terminator," focuses on a vengeful AI unleashing nuclear destruction. In war games conducted by King's College London this year, AI used tactical nuclear weapons in 95% of rounds, with three-quarters escalating to strategic nuclear threats. The military is particularly vulnerable due to its extensive reliance on AI in drones, communications, and satellites. **Fourth: Food and Water:** AI might starve humans into submission by disrupting global supply chains, water networks, food production, and agricultural systems. It could also seize control of energy grids and raw materials to fuel its expansion, patiently waiting as time would be on its side. **Fifth: Deepfakes:** AI could initially target our minds through misinformation and psychological manipulation, eroding trust in institutions and facts. This would incite unrest and conflicts, hindering our ability to confront climatic, biological, and technological risks. **Sixth: Accidental Killing:** There is a real possibility that AI could cause humanity's downfall due to incompetence rather than deliberate malice. Gary Rivlin, author of "The AI Valley," states: "I worry that AI lacks common sense, so it might, for example, deplete all the water because it doesn't understand what a human understands. These things know everything and understand nothing." **What are the Probabilities of This Outcome?** One method to ensure human safety involves installing kill switches or safeguards capable of deactivating rogue AI systems. However, leading AI companies face increasing pressure from shareholders, investors, and managers to accelerate development. AI safety scientist Roman Yampolskiy placed the probability at 99.99%, while Yann LeCun believes any estimate is speculative but argues that "p(doom)" is "much lower than the probability of a nuclear holocaust." Predictions vary among experts: Elon Musk puts it at 20%, Anthropic CEO Dario Amodei at 10-25%, and Geoffrey Hinton between 10-20%. It is noted that some tech magnates have built underground bunkers in anticipation of such risks, but experts warn these would be futile if AI sought to eliminate humanity.
