We've reached the point where we're having to actively degrade the UI and remove useful features of the small scientific database I co-maintain (xronos.ch) to try and reduce the extreme load imposed by (unwanted and politely-asked-to-go-away) AI crawler bots. Conversations about new features have to ask whether it's going to make the crawler load problem worse.
TL;DR: The Florida Supreme Court has introduced a new rule requiring that all legal documents filed in state courts must accurately cite existing legal authorities, aiming to address the issue of AI-generated hallucinations in court filings.
yaknow… the #luddites smashed machines that made usable products.
with a fail rate of as high as 90% in something as simple as search and with sloppped “art” that looks like it was created for a Jehova Witness or Mormon magazine, the only thing #AIslop makes is the mirage of shareholder value.
then there’s the ecocidal damage of data centers. and the theft of public utilities.
“anti-tech extremism” is fascist speak for “protect the mirage of our stocks’ value with cops”.
didn’t i tell you Google is using the #sideloading ban to side-step labor issues? it’s actually worse:
google wants to #aislop and monopolize all #android apps.
so, not only don't they want to deal with developers as contract workers on their appstore ―that’s the whole point of forcing ID, to shoo away the vast majority of developers who aren’t incorporated as businesses. they now want to wipeout said developers and own all the apps. theverge.com/tech/932364/googl…
LinkedIn user hides AI prompt injection in bio to force recruitment spam to be sent in Olde English prose — bots also manipulated to address user as ‘My Lord’
I'm preparing for my class tomorrow where I teach my undergraduates about plagiarism and LLMs. I decided to ask some LLMs questions about myself. One of them led to a funny place...
It's not just you. Scammers, hackers, and other cybercriminals are complaining about “AI shit” flooding platforms where they discuss cyberattacks and other illegal activity.
We've been dealing with the same thing on our invite-only instance. I've been planning to add Cloudflare Turnstile for registrations and maybe a few other mitigations. They're quite sneaky, since the registration reasons often sound totally believable. I've been using this blacklist checker, usually, the email domain is showing up on a blacklist: mxtoolbox.com/blacklists.aspx
Discovered this morning that the spam AI Slop had a human signing up and then inviting the AI Slop as a way to circumvent the Approval sign ups.
Finally managed to ban outright the main culprit email system smtp.oneb.
