To err is human; to really foul things up requires a computer.
— BILL VAUGHAN

To say I am a big fan of computers might be a bit of an understatement. I think they are incredible tools that can provide value from amusement to life-saving research and everything in between, but that doesn’t mean the quote above is wrong. Starting when I was given my first computer in 1983, I have become intimately aware of how horribly wrong things can go when a computer is involved. However, over time, I realized that most issues were caused by entering something incorrectly or not fully thinking through what outcome I wanted. Luckily, it is usually pretty simple to tell when something goes wrong if a) you have a general idea of what the output should look like and b) you are checking the output against that understanding.

I am telling this story because I have been watching the Trump administration generate an enormous amount of really questionable data, likely because they lack both a and b from above. That would be dangerous enough, but it seems that they have fallen in love with the idea of Artificial Intelligence (AI), and when AI is added to ignorance and sloth, chaos ensues.

I am reminded of when my stepdaughter got her first calculator in middle school. No matter what came out, she dutifully wrote it down as the answer. Nothing I could say could convince her to try to pre-estimate the answer and then do a comparison check on the result. Luckily, the first terrible grade and the resulting talk with her teacher got her straightened out. She started seeing the calculator as a tool to help her with math, not a magic box that would do it for her. Her experience isn’t unique; most grow from it and move on. Unfortunately. I don’t see any signs that the Trump administration is as capable of learning and changing as well as a twelve-year-old, and it scares the hell out of me.

There are strong indications that the administration used ChatGPT or something similar to generate their tariff numbers and, even more chillingly, they are also doing so in their review of student visas. Using unexplainable technology that is known to hallucinate for generating lists of people targeted for arrest or tariffs that could crash the world economy is of the utmost stupidity. To do so without verifying the output before using it is criminal negligence. If you wonder why I am sure no verification is happening, see the Rümeysa Ozturk and Suguru Onda situations, where they were flagged for deportation for co-signing an editorial and a fishing license violation, respectively. And what about the tariff impacts on those poor penguins?!?

When employers and organisations see AI as an infallible tool, they may hand over critical decision-making powers to systems not designed for precise, context-sensitive judgements. Tshilidzi Marwala

One of my tendencies, which I see in many of my tech worker peers, is to dismiss it and move on when confronted with something so far out of whack as farcical (surely everyone can see it, right?), unfortunately, with the stakes as high as they are for individual liberties, the economy, and who knows what else, we can’t do that with this administration in power.

As people who have a better-than-average understanding of tech, including how AI works in its current iteration, it falls to us to call out the rancid BS emanating from the White House for what it is. This means explaining to our friends, families, acquaintances, elected representatives, and anyone else who will listen via whatever channels we have at our disposal what is happening and how categorically unfit AI is for life-critical decisions, especially in the hands of fuckwits.

In an era when both the sitting government and a large swath of the population hold ‘experts‘ and ‘knowledge workers‘ in disdain, we have to explain the internet to grandma all over again, but this time to a hostile audience and with people’s lives on the line. Failure.is.not.an.option.

I never said it was going to be easy.