At this point, if you've been paying attention (and even if you haven't been, to be honest), you know that AI is mainstreaming. It's becoming more and more ingrained into our day-to-day lives and is being pushed forward by massive amounts of both governmental and non-governmental R&D funds. What you likely don't know are the true stakes of this competition and who's likely to win.
What we see every day are new consumer products from the big four — Google, Facebook, Amazon, and Microsoft — that are using AI. We see things like Alexa being unveiled, or we hear about self-driving cars. Deep learning systems are being deployed by banks to protect us from fraud (or to shut our cards down when we're on vacation, which happened to me this summer). Sometimes, we hear about the DARPA funding AI research, or we hear about automated drones, but not nearly as frequently.
The latter is where the AI wars are really being fought.
Nations are feverishly working on AI systems to guide the next generations of weapon systems. The country that develops the best AI for these kinds of applications first will have nearly the kind of scalability in weapons systems that we have in software today. And this will be an absolutely overwhelming advantage.
Let's look at aircraft. Today, aircraft are built with the survivability of the pilot as a key design driver — as it should be. If pilot survivability was secondary to, say, weapon load, I don't imagine that you would find many pilots. But imagine if you didn't need to design with the pilot in mind or you didn't even need to train a pilot in the first place — all you needed to do was upload the pilot AI to the aircraft. The aircraft would be much lighter. You wouldn't have all that pesky life support or ejection equipment. Or a cockpit. And the AI could be protected so much better than a pilot. After all, pilots need things like windows and biological support (i.e. using the restroom when you're on long flights). An AI doesn't. And AI won't pass out at high G. So this leads to a smaller, faster, better-armed, and more maneuverable aircraft that cost less and can be built and deployed much more quickly. And imagine the impact on tactics and strategy when the weapons you use to gain battlefield superiority don't involve loss of human life.
Wow. That's kinda a big deal.
Now apply this same technology to submarines (squadrons of underwater sea-planes, anyone?) or land-attack vehicles. With the appropriate communication in place allowing close coordination of these assets, this kind of military would be absolutely unstoppable by conventional means.
The real competitors in the AI wars aren't the companies we think are competing today. They're the nations those companies are from. And those countries know it. Vladamir Putin has called AI development an arms race, while China is pushing to become the prominent hub for AI research worldwide. Likewise, the United States has invested billions in AI and related technologies over the past two decades. Who's winning? Who knows? Arguably, the US and China are neck-and-neck, and neither one is really comfortable with the other getting there first.
This is the real AI war today. Let's hope it doesn't evolve into a real war.