I’m not sure what follows belongs in KJR, and if it does whether it offers anything new and insightful to what’s being published about the subject elsewhere.
Please share your opinion on both fronts, preferably in the Comments.
Thanks. — Bob
# # #
In the game of evolution by natural selection there are no rules. Anything a gene can do to insert more copies of itself in succeeding generations is considered fair play, not that the players have any sense they’re playing a game; not that the concept of “fair” plays any part in their thinking; not that thinking plays any part in most of the players’ lives.
Among the ways of dividing the world into two types of people … no, not “those who divide the world into two types of people and those who don’t …
Where was I? Some of those in leadership roles figure rules are part of the game, and there’s really no point in winning without following them.
That’s in contrast to a different sort of leader — those who consider rules as soft boundaries, to be followed when convenient or when the risk of being caught violating them, multiplied by the penalties likely to be incurred as a result of the violation, are excessive.
For this class of leader, the only rule is that there are no rules. Winning is all that matters.
Which gets us to a subject covered here a couple of weeks ago — the confluence of increasingly sophisticated artificial intelligence and simulation technologies, and their potential for abuse.
Before reading further, take a few minutes to watch a terrifying demonstration of just how easy it now is for a political candidate to, as described last week, “… use this technology to make it appear that their opponent gave a speech encouraging everyone to, say, embrace Satan as their lord and master.”
And thanks to Jon Payton for bringing this to our attention in the Comments.
Nor will this sort of thing be limited to unscrupulous politicians. Does anyone reading these words doubt that some CEO, in pursuit of profits, will put a doctored video on YouTube showing a competitor’s CEO explaining, to his board of directors, “Sure our products kill our customers! Who cares? We can conceal the evidence where no one will ever find it, and in the meantime our profits are much higher than they’d be if we bore the time and expense of making our products safe!”
Easy to make, hard to trace, and even harder to counter with the truth.
Once upon a time our vision of rogue AI depended on robots that autonomously selected human targets to obliterate.
Now? Skynet seems almost utopian. Its threat is physical and tangible.
Where we’re headed is, I think, even more dangerous.
The technology used to create “Deepfake” videos depends on one branch of artificial intelligence technology. Combine it with text generation that writes the script and we’re at the point where AI passes the well-known Turing test.
Reality itself is under siege, and Virtual is winning. Just as counterfeit money devalues real currency, so counterfeit reality devalues actual facts.
We can take limited comfort in knowing that, at least for now, researchers haven’t made AI self-directed. If, for example, a deepfake pornographic video shows up in which a controversial politician appears to have a starring role, we can be confident a human directed tame AIs to create and publicize it.
And here I have to apologize, on two fronts.
The first: KJR’s purpose is to give you ideas you can put to immediate, practical use. This isn’t that.
The second: As the old management adage has it, I’m supposed to provide solutions, not problems.
The best I have in the way of solutions is an AI arms race, where machine-learning AIs tuned to be deepfake detectors become part of our anti-malware standard kit. Or, if you’re a more militant sort, built to engage in deepfake search-and-destroy missions.
That’s in addition to the Shut the ‘Bots Up Act of 2019 I proposed last week, which would limit First Amendment rights to actual human beings.
It’s weak, but it’s the best I have.
How about you?