Anyone who reads this blog must surely know that I consider myself an amateur historian. With that in mind, then, I can’t imagine anyone being surprised that I was intrigued by an article in today’s Washington Post about modern technology and what it’s doing to the past.
In the article, they speak of colorizations of famous black and white photographs, and artificial intelligence either creating famous speeches (e.g., John Donne or Henry V) that were given before we could record them, or — somewhat ominously — speeches that were never delivered, such as Richard Nixon’s speech he was prepared to give in the event of the failure of Apollo 11 or JFK’s planned speech in Dallas.
There’s a good debate worth having here. A few years ago, I saw a fact that was simultaneously new informatio but completely unsurprising: that the terrorist attacks on September 11, 2001 were the single most documented event in world history. This is not surprising because of where it took place and the proliferation of then-available technology. (And changes in technology since then have made other events that haven’t happened yet, even more documentable…)
So it’s an interesting thought experiment: if we could somehow go back in time and instill modern technology into past events, what would we learn about those events that we currently don’t know? Pick any major battle of any major war, for example, and you can get a true, honest picture of things like the number of people on each side, who they were, the injuries they sustained, and so on… We’re always talking about how soldiers return home from a war completely changed because of what they saw on the battlefield and how, after the peace treaties are signed, we say “never again” until the next war comes around. Could technology help make that “never again” a reality without resorting to dishonest portrayals of the event?
So I think that anything that can put a human face on the limitations of then-available technology (like careful colorization of black and white photos) can be a good thing.
Creating the audio of a speech that was never actually delivered is a bit murkier and it definitely tiptoes between the question of what we can do and what we should do.
One thing we certainly don’t want to do is create the audio of one person giving someone else’s speech. I don’t think anyone wants to hear Donald Trump trying to deliver Martin Luther King Jr.’s “I Have a Dream” speech. (And I’m not just saying that as a criticism of Trump; I should think that even his most ardent of supporters would recognize that he and King have/had two completely different speaking styles…)
But when it comes to text of speeches that someone could have delivered under different circumstances, when the text is already known and available? Isn’t it safe to assume that the speaker practiced the speech at least once? Knowing Richard Nixon’s penchant for recording things, the only reason why there is no extant recording of him practicing the delivery of a “doomed Apollo 11 speech” was because he didn’t start recording himself until two years after the fact. If he’d thought of it earlier, there probably would be a recording of him giving it, albeit not to a public audience. And by now it would be available to the public. So what’s the difference between that and a computer simulation of the same speech, given by the same person?
At the very least, a good test of the AI that pieces together a speech like that would be to compare the piecemeal part of the speech, against a recording of a speech that was actually given. Surely someone could get a computer to simulate JFK’s inaugural address based upon the text of it and clips of him speaking, to see how it sounds in comparison to the real thing?
As long as everyone involved is honest about it and what they’re doing, I don’t really see any major problems with this. There will be risks, and growing pains, and people who are less than honest, to be sure, but when all is said and done, I should think and hope that it could create more engagement, especially with amateur historians like me.