Show threadJoanna BrysonMar 1
#NeverTrustAI it's not the kind of thing you can form a relationship with unu.edu/cpr/blog-pos... see also joanna-bryson.blogspot.com/2025/02/gene...

AI & Global Governance: No One...
AI & Global Governance: No One Should Trust AI

It is possible and desirable to engineer Artificial Intelligence for accountability.

United Nations University
Joanna BrysonMar 1
www.404media.co/metas-ai-pat... @[email protected] talking about the future of death, and AI legal agency through the back door. This is particularly terrifying with the stupid "agentic" systems ramping up vulnerability to scams – scams already bankrupt not just individuals but families. #NeverTrustAI

Meta's AI Patent to Simulate D...
Meta's AI Patent to Simulate Dead People Shows the Dangers of 'Spectral Labor'

Researchers say Meta’s patent for simulating dead users could be a “turning point” in “AI resurrections.”

404 Media
Show threadPhelipeFeb 10
e il responsabile è ... (rullo di tamburi) Claude Opus! E il mio collega che non ha verificato a dovere prima di fare il commit e aggiornare il cliente, e quindi anche il sottoscritto come responsabile sviluppo.
Per qualche motivo Claude ha ritenuto consono eliminare una riga che controllava una particolare condizione che ovviamente si è verificata, facendo scattare l'importazione di tutti i record e non solo dei nuovi (ovviamente più volte).
#nevertrustai
Sue CohenJul 15, 2025
"Defense Department to Begin Using Grok, Musk’s Controversial AI Model
A no-bid contract & rushing unproven AI into critical defense systems"
Is it Nazi Grok or Truth Grok?
Jeopardizing American Security with AI?
https://open.substack.com/pub/trumptyrannytracker/p/trump-tyranny-tracker-day-176?r=x9iw&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
#NeverTrustAI
Ivo BathkeJul 5, 2025
Hier lag #ObsIdentify daneben: Das ist wohl nicht Sapygina decemguttata, wie mit 100% vorgeschlagen, sondern Sapyga quinquepunctata, erkennbar an dem roten Abdomen. #nevertrustai Btw meine 1202te Specie auf naturgucker.de
Show threadMatt BrailsfordNov 1, 2023
@owaincodes WTH, lol. I didn’t actually check if the summary was right. Looks like ChatGPT can make stuff up when you ask it to TLDR a URL. Good to know for future 👍🏻 #TeethingProblems #NeverTrustAI