Are Glorified Autocompletes Going to War?

Ever since the topic of Artificial Intelligence (AI) became "hot" again in late 2022 we have been inundated with coverage of any and every hint of an imminent (or not so imminent) development in the field, as well as speculation about what it all might mean.

Those who are bullish, of course, get more attention--the idea that "Something big is happening" more interesting to most than explanation as to why that supposedly big happening is really "Much ado about nothing." Still, watching this field these past few decades I have consistently seen the hype far, far outrun the reality--certainly to go by what Kurzweil predicted in 1999, the possibility that Frey and Osborne argued in 2013, what Musk promised over and over and over again. Meanwhile Robert Gordon's skeptical views from back in 2000 have never been far from my thinking, and with them the implication I still argued this year--namely that any real artificial intelligence revolution will come when engineers advance AI performance in those key areas of perception of, navigation through and dextrous manipulation of the physical world to the point at which they can reliably perform physical tasks like moving about and handling objects as well as an adult human. And even as AI programs astound observers with their capacity to generate words and images, that prospect still seems remote.

Still, amid a profound remilitarization of international relations the world's armed forces--which necessarily take the view that they cannot afford to dismiss what is only "probably" hype just in case there really is something to it (and frankly, have historically been very susceptible to mere hype themselves)--the longstanding interest of those forces now seems to have a sharper edge as what may have seemed like unrealistic aspirations a few years ago are now being tested out, with this the case with such systems as fighter aircraft. Where a short time ago it seemed as if air forces were drawing back from their earlier, extravagant, expectations for the next, sixth, generation of jet fighters, the U.S. Air Force is now putting the "X-62" through its paces--an F-16D equipped with AI (a fourth-generation fighter with a sixth-generation control system) that has, reportedly, permitted it to take on a human pilot in a mock battle.

Is this the beginning of a process that will go all the way to a new "Revolution in Military Affairs," or is this line of development going to fizzle out the way the promise of so many technologies has in the past?

I suspect that we will get a clue as to the chances of that in what happens with AI in the civilian economy, where the stakes are apt to be less great, the tolerances not so small, as they are in high-technology warfare--and it must be admitted, even mainstream business news coverage is increasingly using the word "bubble" to describe the situation.

If they are right about that it would be far from the first time that defense planners "jumped the gun" in thinking that a still-developing technology was ready for the battlefield.