Did We Learn Nothing From ‘WarGames’?
Perhaps you’ve been reading what passes for the news over the past week or so: The government wants the machines to take over, they will soon have control of U.S. military operations, and we’re all going to die.
We’ll take a few steps back. The tech company Anthropic, best known as the “frontier lab” behind the kinder, friendlier, far-less-likely-to-spew-racist-garbage chatbot Claude, is currently engaged in a fight with the Department of Defense. There are, shall we say, some differences of opinion over how the current administration could — and should — use Anthropic’s A.I. system. Specifically, CEO Dario Amodei is wringing his hands over the potential lack of guardrails once the system is handed over. Amodei released a statement in which he noted that the company, which has a contract with the U.S. government and has deployed its artificial-intelligence models across several classified federal networks, has “never raised objections to particular military operations nor attempted to limit use of our technology in an ad hoc manner. However, in a narrow set of cases, we believe AI can undermine, rather than defend, democratic values. Some uses are also simply outside the bounds of what today’s technology can safely and reliably do.”
Those uses, per Amodei, involve mass surveillance of Americans, notably in ways that extend far past trying to find missing pets, and “fully autonomous weapons.” In plain speak, the term refers to AI systems having complete control over the targeting and usage of the military’s complete arsenal, including thermonuclear weaponry, and keeping human beings — remember them? — out of the loop. “Fully autonomous weapons cannot be relied upon to exercise the critical judgment that our highly trained, professional troops exhibit every day,” the memo states. “We will not knowingly provide a product that puts America’s warfighters and civilians at risk. We have offered to work directly with the Department of War on R&D to improve the reliability of these systems, but they have not accepted this offer.”
Editor’s picks
Those last six words explain where we are right now. Anthropic wants to amend its current contracts in order to avoid, y’know, a possible mecha-apocalypse. Secretary of Defense Pete Hegseth has stated that the U.S. government should be able to do whatever the hell it wants with Anthropic’s software on the battlefield, Armageddon be damned. He is willing to declare the company a “supply chain risk,” and cancel any and all contracts. The president naturally weighed in as well. Many critics, political pundits, and wags asked, somewhat rhetorically, whether anyone in the administration recalls how handing the responsibility of nuclear weaponry over to machines turned out in the Terminator movies.
We’re curious, however, to know if the administration remembers another popular film from that same time period: WarGames.
As any die-hard fan of Reagan-era cinema will tell you, the 1983 hit revolved around a Seattle computer nerd named David Lightman, played by Matthew Broderick as dorky but in a sorta cute, pre-Ferris Bueller way. He’s a wiseass in class, and a whiz at hacking into the school’s network and changing his grades. When he tries to help out his fellow failing student Jennifer (Ally Sheedy) by switching her F to an A, she balks. But the young woman is impressed by David’s brains and his cojones. You can practically smell the pheromones emanating off these two teens.
Meanwhile, over at the headquarters of the North American Aerospace Defense Command (NORAD), everyone’s losing their shit over a recent training exercise. We’ve already seen an employee refuse to facilitate releasing ICBMs out of a U.S. missile silo during this drill; not even his partner pulling a gun on him will get him to potentially kill millions of people. (Yes, the guy holding the pistol and barking “Turn your key, sir!” is indeed a baby-faced Michael Madsen. Even then, the future Mr. Blonde was genuinely cold-blooded.) Apparently, this reluctant mass murderer was not the only one failing to comply. Nearly a quarter of the men manning those remote control centers refused to turn their keys as well.
Luckily — “luckily” — one guy has a solution. He’s Dr. John McKittrick; given that’s he played by the decade’s go-to a-hole Dabney Coleman, you can immediately tell he’s a piece of work. McKittrick wants to outsource control of America’s nuclear arsenal to a supercomputer known as the War Operation Plan Response, or WOPR for short. This machine does nothing but run simulated plans of attack and/or responses to attacks 24/7. “It spends all of its time thinking about World War III,” he boasts (which is probably an accurate description of Hegseth as well). McKittrick is aware that the commander in chief is technically the one to dictate strategy in terms of international conflicts, but he’s convinced “the president will probably follow that war plan … once he makes that decision.”
Long story short, David accidentally hacks into WOPR’s network, and faced with a menu of games that includes chess, checkers, poker, Fighter Combat, Air-to-Ground Actions, and Theaterwide Biotoxic and Chemical Warfare, he picks the most intriguing title: Global Thermonuclear War. The enterprising youngster initiates a war game that the computer keeps playing even after he’s logged off, with very real stakes. He and Jennifer are forced to go on the run as the feds chase after them and must eventually trick WOPR into a stalemate before everything goes boom.
The storyline between Broderick and Sheedy was something that new director John Badham was keen to put in the forefront. (Badham took over the project after the original filmmaker, Martin Brest, was let go; the firing did free Brest to take on a new film, however, some fish-out-of-water story called Beverly Hills Cop.) And while most moviegoers of a certain age easily think of WOPR’s computerized voice asking “Shall we play a game?” whenever someone brings up WarGames now, the movie ended up becoming a huge hit less because of its topicality and more because it was tapping into a steadily growing demographic of adolescent audiences. The whole thing plays like a cross between a 1970s conspiracy thriller and the type of movie that would soon become John Hughes’ specialty. Call it Teen Days of the Condor.
But the concept that a computerized system controlling U.S. military actions could be compromised had spooked the powers that be. Shortly after WarGames was screened at the White House in June 1983, Ronald Reagan asked during a national security meeting if this sort of fanciful idea was possible. Several advisers humored the president’s request that they look into it. General John W. Vessey Jr., then the chairman of the Joint Chiefs of Staff, came back a week later and told Reagan that “Mr. President, the problem is much worse than you think.”
It turns out that a number of scientists and policy wonks had been studying this very idea for many years prior to WarGames, but it took the movie to make Reagan inquire and bring what had been niche hypotheticals into the broader political conversation. That teen flick would affect how the U.S. government dealt with cybersecurity and the notion of cyberwarfare for generations. Presented with a fictional worst-case scenario, the real-life counterparts in the military brass and inside NORAD began to look seriously into safeguards. They realized, like WOPR does in the film, that they were playing a zero-sum game. As did the general public — during one of the special features on the DVD release of WarGames, Badham noted that when the supercomputer announced at the end that “the only winning move is not to play,” audiences stood up and cheered. It might just be a good story, but we’ll take apocryphal over apocalyptic any day.
Which brings us back to Anthropic, and Hegseth, and our current moment. The AI company is demanding that basic morality and accountability be a factor in any use of their software, or they will refuse to license it. Hegseth countered with saying that nobody could tell him or the administration what to do, and he’d cancel all their contracts and just go find another AI company without a moral compass that won’t put up a fuss. The staring contest continues, even as the administration gets itself involved in a war.
Have they learned nothing from WarGames? Have we learned nothing from the movie? There is a sense that Hegseth and his cohort, if any of them have seen the film at all, stop watching it after the first 15 minutes or so. That would have been just long enough to see Dabney Coleman’s character say that expressing hesitation or remorse over millions of deaths was “headshrinker horseshit”; to hear someone note that “this trillion-dollar hardware is at the mercy of those men with the little brass keys,” and have Coleman reply, “… whose only problem is that they’re human beings”; and for him to essentially conclude that the only way to ensure victory in warfare was: “I think we oughta take the men out of the loop.” Judging from what we’re seeing happen at the Pentagon now, these folks never got around to the last act, where Coleman sees that eliminating the human factor in warfare means potentially eliminating the human factor, period. They missed the whole part about the only winning move is not to play with this dangerous idea at all.
