"We're bombing Iran to bring back Jesus"

My brother in Christ, He's already here.

He is the family of refugees you deported for failing to have their papers with them at a Walmart. (Leviticus 19:33-34)

He is the hungry kid you denied food by cutting his SNAP benefits. (Matthew 25:35)

He is the sick grandmother who can no longer afford her medication because you let DOGE plunder her social security. (Deuteronomy 10:18-19)

He is the homeless person you arrested simply for sleeping outside when he had no other option. (Zechariah 7:10)

He is the trans woman you just convicted of a "sex crime" for daring to exist in public. (Matthew 19:12)

He is the poor, the disadvantaged, the abandoned, the elderly, the sick, the weary, and the downtrodden. And you - yes YOU SPECIFICALLY - will be judged by Him by how you treated Him here on earth. (Matthew 25:45)

This is the script of my national radio report yesterday on the controversy over Grammarly's "Expert Review" AI personas of actual people, that were created without permission. As always there may have been minor wording variations from this script as I presented this report live on air.

- - -

Yeah it seems like every time we explore an outrageous issue regarding AI, we want to say, well it can't get worse than this one, but then a bit of time goes by and whammo there's another one that just demonstrates that when it comes to these Big Tech AI CEOs there's just no bottom, and they seem to live in a world completely different from the one most of us inhabit.

So it's important to remember that by and large the entire generative AI Large Language Model industry is based on the mass appropriation of data, of content, by these powerful firms used to train their AIs, typically without any direct compensation offered to the sources nor permission asked in advance. And this has taken place on a colossal scale, mind-numbingly enormous and apparently largely driven by the view by these firms that they will never be held to account for what they do so long as they continue to have friends in high political places.

And of course we've talked over time about a whole range of staggeringly awful aspects of these AI systems including how they're trained, the kinds of misinformation that have become so associated with them, and how generally these AI firms don't want to take responsibility for damages done to anyone who uses those systems -- which are becoming increasingly difficult to avoid using.

Well, today we have yet another example that again demonstrates that the raw hubris frequently associated with these AI firms seems to be effectively unlimited. You may know of the Grammarly website that is used to provide spelling, punctuation, and grammar checking and other services related to written materials, that they've increasingly pitched as an AI-powered service. Recently they changed the name of the firm itself from Grammarly to something a bit less modest. The company is now called SUPERHUMAN. Never expect modesty when it come to Big Tech.

So, Superhuman slash Grammarly came up with what they thought was a sure winner AI feature that they call "Expert Review". They created virtual AI personas of actual people that they considered to be notable, both alive and deceased, so that users could have their writing "critiqued" by these virtual creations. So they included famous scientists, writers, columnists, you get the idea. And then the AI virtual person would supposedly offer opinions and suggestions, ostensibly as the real person would, or in the case of deceased persons, supposedly would have if they were still alive.

And of course you'd expect that Superhuman paid these actual people -- at least the living ones -- handsomely to participate in this, right? WELL you know where this is going, they didn't pay these people handsomely, they didn't pay them anything. But at least they asked for permission before using their personas, right? NO, of course not, permission wasn't asked. Well, this has all blown up into a massive controversy -- a class action case has apparently already been filed.

Many of the persons whom Superhuman simulated were understandably very upset when they found out about this, not only because this was done without payment or their permission but because of the obvious risks of reputational damage in such situations from whatever those virtual AI personas spout. So in response Superhuman set up an opt-out email address -- totally inadequate -- and then as the controversy continued took the service down at least for now, saying that they they had "missed the mark".

This is classic Big Tech "move fast and break things" -- "plow ahead without permission and apologize later if you have to" BIG HUBRIS thinking. Yet another example of why society should demand that the AI industry be subject to very stringent regulations no matter how many politicians in either party try to protect those AI firms.

This is a matter of basic humanity that transcends technology and politics. We need strong AI regulations. We needed strong AI regulations before now, but at the very least we need them immediately, before even more AI abominations are spewed from those Big Tech firms that seem to care plenty about their bottom lines, but not at all about us.

- - -

L

The three skills with a lot less overlap than you’d expect:

1. Ability to code.
2. Ability to perform well in a coding interview.
3. Ability to validate code.