AI, Gold and Nuclear War

So-called artificial intelligence (AI) is taking the world by storm. Meanwhile, gold has shot up like a rocket over the past couple of months.

In mid-February, gold was trading at $1,990. Two months later, gold is trading above $2,400 — a $410 gain in just two months.

So here’s a question:

Is there a connection between AI and gold? It seems like an odd question. But as it turns out, the answer is yes. And surprisingly, there has been for decades. It involves the Cold War between the U.S. and the Soviet Union.

In the early 1980s, the KGB was deeply concerned about the possibility of a nuclear first strike by the United States. At the time, Yuri Andropov was head of the KGB.

Andropov’s fear of a nuclear first strike by the U.S. was based in part on the 1980 election of Ronald Reagan and Reagan’s plan to install Pershing II intermediate-range missiles in Europe.

Those missiles could be armed with nuclear warheads and could strike the Soviet Union within minutes of being launched. This put Soviet nuclear forces on a hair-trigger alert. They adopted a “launch on warning” posture.

This means that as soon as credible evidence of a planned first strike was discovered, the Soviet Union would launch its own first strike to avoid destruction of its forces.

The irony was that the U.S. had no actual plans to launch a first strike, but the Soviet Union didn’t know that. Reagan’s speeches about the “evil empire” did nothing to calm Soviet concerns.

AI and Nuclear Readiness

In response, the Soviets developed a primitive (by today’s standards) AI system called VRYAN. That’s a Russian acronym for: sudden nuclear missile attack.

VRYAN took about 40,000 military, economic and political inputs and computed the relative strength of the Soviet Union compared with the United States expressed as a percentage output. The model used a value of 100% for equivalence of the USSR to the U.S.

The Soviet leadership was comfortable that the U.S. would not launch a nuclear first strike if the USSR could maintain a value of 60%, although they viewed 70% as providing a more comfortable margin.

A VRYAN output of 40% was considered the critical threshold at which the U.S. might feel it could launch a first strike with acceptable risk that the Soviets would not be able to mount a successful second strike.

VRYAN output values were in steady decline in the dangerous period from 1981–1984 (in 1984, the VRYAN output had declined to 45%).

The VRYAN AI system relied on by the KGB and the Soviet Politburo was an important factor in the Soviet decision in 1981 to vastly increase intelligence collections aimed at detecting U.S. preparations for a first strike.

Close Call

This intelligence collection effort was complicated to the point of extreme danger by the fact that the U.S. and NATO were conducting a war game in late 1983, code-named Able Archer 83. This war game was to practice a nuclear strike on the Soviet Union.

It turned out that the U.S. was rehearsing a nuclear first strike at the same exact time that the KGB was looking for evidence of a nuclear first strike. Able Archer 83 provided the KGB with more than enough reason to suspect the U.S. was indeed preparing for a first strike under cover of a war game.

VRYAN’s AI output on relative U.S. strength was compounded by massive U.S. intelligence failures regarding Soviet intentions. U.S. intelligence analysts assumed that the future would resemble the past, and that Soviet alerts were really propaganda designed to halt the U.S. deployment of Pershing II intermediate-range nuclear missiles in Europe.

U.S. intelligence analysts were also guilty of what’s called mirror imaging: the belief that because you know your own intentions, your opponents must share your view. In this case, the U.S. assumed that because they had no intention to launch a first strike, the Soviets must have understood that intention and would therefore have no cause for concern.

In fact, the Soviets had the opposite view based in part on VRYAN AI output.

The world came extremely close to World War III and a nuclear holocaust as a result of this sequence of events and misperception of intentions. It was only when one U.S. general decided not to escalate in the face of Soviet first strike preparations that both sides deescalated, and the crisis eventually receded.

The information above wasn’t fully understood by either side at the time of the escalation. On the U.S. side, it wasn’t until the 1990 publication of a study entitled The Soviet War Scare by the President’s Foreign Intelligence Advisory Board (PFIAB) that something like the full story was revealed.

Nuclear War Threats: Good For Gold

This study was originally classified above TOP SECRET. (Most citizens assume that TOP SECRET is the highest level of classification. But there are secret access codes that limit circulation of certain documents even among those cleared with TOP SECRET access.

In the case of The Soviet War Scare, those restrictions had the code names UMBRA, GAMMA, ININTEL, NOFORN, NOCONTRACT, ORCON. I can’t discuss my own TOP SECRET clearances, but I can inform you that very few intelligence operatives would have been able to view the PFIAB report based on those restrictions.

So what does all this have to do with gold?

Buried inside The Soviet War Scare was this passage about the U.S. assessment of KGB collection requirements related to a potential nuclear war:

VRYAN Collection Requirements – Throughout the early 1980s, VRYAN requirements were the No. 1 (and urgent) collection priority for Soviet intelligence… They were tasked to collect:… monitoring of the flow of money and gold on Wall Street as well as the movement of high-grade jewelry, collections of rare paintings and similar items. (This was regarded as useful geostrategic information.) (Emphasis added)

And there it is! The U.S. assessed that the KGB tracked the movement of gold as a leading indicator of nuclear attack.

I didn’t find this completely surprising. From 2004–2010, I was co-director of a CIA effort called Project Prophesy that looked at capital markets activity as an early warning of an enemy attack.

Gold was one of the valuable assets that was on our list of items to track. The idea was that if a general or political leader had advance information about an attack, they’d convert their wealth to gold in safekeeping in order to financially survive the fallout.

The bottom line is that this intelligence reporting and AI system are not ancient history. Today, the world is closer to nuclear war than at any time since the Able Archer scare in 1983. Gold is once again on the move, having risen from $1,830 per ounce on Oct. 5, 2023, to over $2,400 today. That’s a 31% gain in six months.

Is this a coincidence? Hardly. A close correlation of huge gains in gold with serious threats of nuclear war is exactly what one should expect.

Unfortunately, those threats of nuclear war are not going away soon. One need only look at the Iranian attack on Israel this past weekend and the possibilities of further escalation.

There are also situations in Ukraine, Russia, NATO, Gaza, the Red Sea and the Suez Canal revealing that the world is a more dangerous place than it has been for decades.

That’s bad news for the world but good news for gold investors. The rally we’ve seen in the past six months is just getting started.