I Too Am Artificially Intelligent

Too much AI bullshit. An airline trusted AI too much and now has to pay up. Reddit decides to sell their user content to AI companies for peanuts. OpenAI missed the point of Oppenheimer. Russia likely killed a prominent dissenter.

Endeavorance

Feb 19, 2024 — 5 min read

Listen to the podcast version here.

Programming Note

Quick scheduling note, there will be no Stuff Keeps Happening this Thursday, Feb 22nd. SKH will return Monday, February 26th. Thanks for understanding!

Airline Chatbot Stands In Solidarity

Weird! Maybe racing to roll out AI customer assistant agents for every single purpose is a bad fucking idea. Who would have thought?

Air Canada—much like virtually ever other corporation—launched a chatbot AI agent, likely powered by ChatGPT in the backend. Everyone is doing this right now, and its fucking dumb becauuuuseeee….

Jake Moffatt's grandmother died. My condolences Jake, sincerely. Upon hearing the news, Jake tried to book a trip with Air Canada, asking the AI chatbot how bereavement rates worked. The chatbot then just made up some complete bullshit, which is what AI chatbots do. But the AI was positioned as a customer support rep, so Jake took it as policy, rightfully so. And so, Jake booked a flight.

Air Canada later refused to honor the policy their own AI produced. The chatbot told Jake he could request a refund. Air Canada said no. So he sued.

Their defense in court? "The AI doesn't represent us." The judges response? "Then why the FUCK do you have an AI customer support rep on your website?"

Jake was awarded a partial refund, and Air Canada's AI chatbot is no longer online. So, lol.

Reddit Sells Your Data To Robots For Peanuts

After their controversy last year, Reddit is seemingly interested in reopening the wound to specifically inject salt into it.

Last year, Reddit decided that they no longer give a shit about the people who use their platform. They did a 180 on the standing policy of their company for the entirety of their history, shutting out third party clients and charging exorbitant fees to use their API, citing Elon Musk's changes to Twitter as an inspiration. When users and volunteer moderation staff complained, the CEO of Reddit called them "Landed Gentry."

The reason for this was so they could have more control over your data so they can sell it for more money. And now that's literally happening.

Reddit signed a deal with an "unnamed AI company," selling their user-generated data for the AI to train on for a mere $60 Million.

If you have posted on Reddit, your content has been sold to an AI company so an AI can impersonate your work without your consent or any payment to you, while Reddit has valued that at $60 Million. I cannot express how little money that is in the tech world. I legitimately thought that figure was a joke the first time I read it.

So that's where Reddit is at. Where once it was founded by a person who was so focused on an ethical, open internet that he literally killed himself to avoid jailtime after a conviction for his activism. Now, the other founder is shitting on his grave while kissing Elon Musk's ass.

Fuck Spez.

Nobody Asked For This

Last up for our AI bullshit roundup, OpenAI's Sora.

Sora is OpenAI's latest attempt at pushing my blood pressure into medically inadvisable levels. It is an AI model capable of taking text input prompts and generating "realistic" video output. Granted, the video looks really good at times, but other times it glitches the hell out and is obviously fake.

BUT, that doesn't really matter now, does it? AI image generation can produce extremely realistic results, but even results with obviously broken shit can trick people who don't otherwise pay attention to the state of AI image generation, which I assume is basically like, everyone. ESPECIALLY older folks who mainline Facebook and shit.

The destructive potential of Sora is obvious, and the immediate jumping to the defense of its existence by weird chuds is predictable. Nobody asked for this. This does not solve a human problem. Just because it can exist doesn't mean it needs to exist.

I've seen the argument that its better for us to make it than someone else, to which I ask: should we be dropping innovative chemical weapons on the entire planet because someone else may, in theory, do that? Because that's what this is but for abuse, exploitation, creative displacement, and disinformation.

Why is this the technology we must defend and support? This IS the national security threat.

Nobody asked for this.

Countless women and children are about to have photo-realistic pornography made of them. Scammers will generate AI videos that look like surveillance cameras proving you did some crime, and will turn it in to cops unless you pay them. Governments will generate propaganda at rates we've never seen before.

All the while, the world has eaten up the rhetoric of the AI savior, and a few kingpins of the tech bro world are being enriched beyond comprehension.

Nobody asked for this.

Alexei Navalny Has Died

Okay how about a break from AI? Oh, fuck. Maybe we should go back to AI stuff. Oh, well.

Alexei Navalny, notable Russian opposition activist, has died in prison. Yes, he was almost definitely assassinated by the Russian government, who previously tried and failed to assassinate him in 2020.

In 2020, he was flown to Germany to receive care after being poisoned with the "Novichok" nerve agent that Russia has used to kill quite a few people around the world now. He recovered, only to ultimately be arrested and jailed when he returned home. At the time of his sentencing, he essentially acknowledged it was a death sentence.

Here's a quote. I cannot fathom being in this position.

If they decide to kill me, it means that we are incredibly strong. We need to utilize this power to not give up, to remember we are a huge power that is being oppressed by these bad dudes.