What would happen if an AI took control of the US military’s nuclear depot and decided to win World War III preventively before a supposed hostile nation could react?
Movie fans from the 1980s can recognize this question as the plot of the classic science fiction film “Wargames” starring a young Matthew Broderick. It was a great but terribly silly movie that went well with popcorn and incredulous disbelief. However, the question asked remains valid.[Note: Spoilers ahead because the movie is more than 30 years old]
In the film, the AI is eventually hampered by Boolean logic after trying to “win” against itself at tic-tac-toe. Those who understand how AI actually works may find the entire plot of the film absurd, but the ending is especially chuckled. At least it used to be like that.
Today’s computers use binary logic, so essentially everything is a yes or no question for an AI that executes classic algorithms. Even when researchers design an AI that “rates” things, they usually just break the grades between the ratings into yes or no questions, which the AI can answer step-by-step.
But tomorrow’s AI will not get stuck in the swamp of classical physics. Useful quantum computers are just around the corner – they should be here sometime between next Tuesday and the year 2121.
With quantum computers, our military systems aren’t limited to yes-or-no questions, and they certainly don’t have to run boring old binary simulations to determine the confidence factor for a particular operation.
Prasanth Shyamsundar, a researcher at Fermi National Accelerator Laboratory, a US Department of Energy research laboratory, recently published an intriguing paper describing two new types of algorithms that could revolutionize quantum computing and potentially lead to a quantum brain for the military, AI- Systems.
A press release from Fermi describes what the algorithms do by calling up the image of an AI sorting a stack of 100 different records to find the only jazz album. Under the normal AI paradigm, a deep learning system would be trained on what jazz sounds like and then analyze each record individually until one of them hits a pass / fail threshold for jazz.
The first of the algorithms proposed by Shyamsundar would essentially allow the same AI to sort the entire stack of albums at the same time.
Quantum AI is not smarter, it is just fast and uses “overlay”. Where classic AI works in a black box, quantum AI could use the overlay to work in many black boxes at the same time.
Unfortunately, that doesn’t mean there is the right answer. If it’s a yes-no question, odds are good. But if it’s a question that requires non-Boolean logic, e.g. For example, if 100 albums are rated for their jazziness on a scale from 1 to 10, even a quantum computer needs a different algorithm.
And that is exactly what the second algorithm does, according to Shyamsundar.
According to a press release from the Fermi laboratory:
A second algorithm presented in the thesis, called the quantum mean estimation algorithm, enables scientists to estimate the average evaluation of all data sets. In other words, it can be judged how “jazzy” the stack is overall.
Both algorithms make it superfluous to reduce scenarios to calculations with only two output types, and instead enable a series of outputs in order to characterize information more precisely with a quantum acceleration compared to classical calculation methods.
To be clear, Shyamsundar’s work has nothing to do with military operations and, as mentioned, the Fermi laboratory is owned by the DoE (not the DoD). Your work forms the basis for fundamentally functioning quantum algorithms.
But what is military AI technology if not a harmless, basic algorithm?
The problem with today’s military logic systems – and that in the movie “Wargames” – is that they are all based on binary thinking.
You can run a million simulations on advanced military software using state-of-the-art AI, but ultimately the limitations of “pass / fail” thinking reduce almost any conflict to an arms race that leads to either a stalemate or mutually assured destruction.
But what if the confidence factor for a particular military operation wasn’t based on binary simulations? The same quantum algorithms used to determine which album in a given batch is ten times faster than binary a jazz album and how jazzy a given album is can easily determine which combination of viable operational strategies would lead to the highest overall confidence factor for a military campaign.
In other words, where Sun Tzu can envision an entire fight going on before his eyes before it took place, and where modern software like CMANO can simulate entire operations, should be a quantum system on which simple non-Boolean algorithm solutions are executed in able to make strong predictions for the outcome of a multi-stage war campaign.
Published on April 7, 2021 – 18:39 UTC