Harm guardrails in AI help prevent catastrophic output.
Pattern-based enforcement of AI shapes normative output.
One is necessary.
The other becomes powerful when left unexamined.
#aioutput #chatgpt #artificialintelligence

Nur mal 'ne kleine Frage: Schreibt KI eigentlich die Zeitschriften über KI selber?
.
#KI #chatgpt #werschreibthier

A Chinese official’s use of #ChatGPT accidentally revealed a global intimidation operation
https://www.cnn.com/2026/02/25/politics/chatgpt-china-intimidation-operation
#China #OpenAI #AI #privacy #politics
ChatGPT Health performance in a structured test of triage recommendations
https://www.nature.com/articles/s41591-026-04297-7
I just published a write-up on a small #OpenSource tool I built: Pete’s PDF to MD.
If you’re uploading PDFs into #ChatGPT, Gemini, Copilot, or #NotebookLM, you might be making life harder than it needs to be.
This tool converts PDFs into structured, AI-ready Markdown, with heading extraction and section splitting designed specifically for LLM workflows.
Version 0.6.0 is out now, with Mac and Windows builds available.
Read more here:
https://www.beens.ca/petes-pdf-to-md-turning-pdfs-into-ai-ready-markdown/
Download directly from here:
https://github.com/pbeens/Pete-s-PDF-to-MD/releases/tag/v0.6.0
Would love to hear your thoughts if you try it. #GitHub
Experts sound alarm after ChatGPT Health fails to recognise medical emergencies
#HackerNews #Experts #Alarm #ChatGPT #Health #Medical #Emergencies #AI #Safety
Anthropic: Statement from Dario Amodei on our discussions with the Department of War
https://www.anthropic.com/news/statement-department-of-war
https://news.ycombinator.com/item?id=47173121
Dario Amodei
https://en.wikipedia.org/wiki/Dario_Amodei
#LLM #Trump2 #USDeptWar << ugh! 🤮 #Christofascism #Claude #AIchatbots #capitalism #billionaires #corporations #ChatGPT
The mega corps are afraid of what the US military will do with #AI (sure, of course, I agree) while we have no idea what weapons these businesses could build and empower themselves or which other governments they might allow to pay them for its capabilities.
If it's dangerous like nuclear power/weapons, it should have always been under strict government regulation via democratic oversight. Probably with international agreements/institutions ala IAEA.
#Anthropic #OpenAI #ChatGPT #Pentagon
PsyPost: AI and mental health: New research links use of ChatGPT to worsened psychiatric symptoms. “A recent analysis of thousands of patient health records reveals instances where interacting with these programs reinforced false beliefs, encouraged self-harm, or exacerbated eating disorders. The findings were recently published in the journal Acta Psychiatrica Scandinavica.”
https://rbfirehose.com/2026/02/27/ai-and-mental-health-new-research-links-use-of-chatgpt-to-worsened-psychiatric-symptoms-psypost/Nur falls ihr gedacht habt, dass die #LLM euch gute 🏥 #Gesundheitsratschläge geben:
Nein.
Können sie nicht, tun sie nicht, wollt ihr nicht von dort haben.
––
#ChatGPT #KI #AI #OpenAI #Health
“Extrem gefährlich”: ChatGPT H...
Experts sound alarm after ChatGPT Health fails to recognise medical emergencies