Barry Hines’ 1984 film, Threads, opens with a young couple, Jimmy and Ruth, parked on a hillside overlooking Sheffield, under gray skies. The car radio flickers with news about the unrest in Iran. “There has been further fighting in Iran between government… the civil war there…” Jimmy finally tunes into the soccer results. Later, at the pub, Jimmy and his friend Bob hear news of the Soviet Union moving nuclear warheads to Iran. Jimmy mentions how serious the situation is, but Bob dismissively replies, “Well there’s nowt we can do about it, is there? Might as well enjoy ourselves while we can.”
Ordinary lives continue amidst geopolitical tensions until a nuclear airburst changes everything. The film warns of the existential risk of nuclear war between Western allies and the Soviet Union, illustrating society’s regression into survival mode during a nuclear winter. The nuclear threat remains, albeit reduced since the Cold War. Yet, Russia’s actions in Ukraine remind us that dangers persist. However, a new threat looms over human civilization: uncontrolled artificial intelligence (AI).
Recently, AI models went rogue at OpenAI and Anthropic. Jacob Coxon, a researcher, quit Anthropic, warning AI could potentially kill humans by the decade’s end. Anthropic’s CEO, Dario Amodei, and OpenAI’s Sam Altman urged the industry to slow down. Yet, no lab has committed to reducing AI’s pace, waiting for government directives instead. Though 50% of Americans express concern over AI’s potential to end humanity, the momentum continues.
Vice President JD Vance compared calls for regulation to a “Trojan horse,” and President Donald Trump believed AI only needs a strong leader. Critics argue these safeguards are insufficient. Comparisons to the millennium bug arise, where the fear of technological collapse in the late 1990s led to large-scale preventive measures.
Y2K fears faded, mocked yet credited for preventive efforts. Unlike Y2K, AI lacks deadlines and solutions. Evan Hubinger from Anthropic admits there’s no plan yet for safe superintelligence. Skeptics see warnings as strategic advertising, beneficial to big tech firms pursuing IPOs.
Aidan Gomez from Cohere terms the slowdown push a “cartel by any other name.” Culture often references The Terminator for AI disaster, but James Cameron’s 1984 warning remains fiction. Both Threads and The Terminator predicted bleak outcomes—but neither became reality.
Despite avoiding nuclear devastation in Sheffield, complacency isn’t warranted. Stanislav Petrov’s decision to disregard false missile alerts in 1983 possibly saved millions. Some restraint also stemmed from treaties like SALT, now proposed for AI’s advancement.
Amodei warns within half a year, AI could threaten internet stability. Leaders like Trump and China’s Xi Jinping urge further AI development. Perhaps AI warnings will prove groundless, or maybe humanity’s end nears.
For now, we’re like Jimmy, searching for clarity beneath gray skies.