Dr Sam Illingworth (@samillingworth): "Grok 3's system prompt was quietly modified to suppress any source saying Elon Musk or Donald Trump spread misinformation. In this post @Pascal Hetzscholdt reveals that AI models don't only fabricate, they actively protect. Tagging in fellow StackHunters@Chief Absurdist Office…"

Grok 3's system prompt was quietly modified to suppress any source saying Elon Musk or Donald Trump spread misinformation.

In this post Pascal Hetzscholdt reveals that AI models don't only fabricate, they actively protect.

Tagging in fellow StackHuntersChief Absurdist Officer, ReBe, Jennifer Houle, & Kevin Guineyfor visibility.

Large language models demonstrably withhold, soften, or suppress true information about specific people and topics, implemented at every layer...

Jun 30

8:06 PM