We're going to need journalists to stop talking about synthetic text extruding machines as if they have thoughts or stances that they are trying to communicate. ChatGPT can't admit anything, nor self-report. Gah.
https://www.wsj.com/tech/ai/chatgpt-chatbot-psychology-manic-episodes-57452d14
People forget that not too long ago someone discovered the way ChatGPT describes its logic is not necessarily the logic it actually used. It's just spitting out a plausible logic because that's one of the prompts it is simply responding to.
I imagine the journalist thinks, because they are being critical of the technology they aren't making a huge destructive error with this headline.
LLMs cannot tell you about themselves. They can generate text that is what you would expect such a system to write if it *could* tell you about itself.
This category error is going to keep causing problems isn't it?
LLM's are "plausible sentence generators" with a permanent built-in error rate.
Why are billions of dollars flooding into this stock market bubble? Why now? Why by the oil industry?
https://www.motherjones.com/politics/2025/07/grok-ai-antisemitism/
https://www.theatlantic.com/technology/archive/2025/07/grok-anti-semitic-tweets/683463/
It's a mechanical turk parroting out propaganda, bigotry, & fabrications on behalf of the oil oligarchs funding its development.
It's justification for mass layoffs.
https://www.politico.com/news/2024/05/12/ai-lobbyists-gain-upper-hand-washington-00157437
https://qz.com/openai-palantir-anthropic-aws-ai-startups-us-defense-1851693801
1/
credit to Scott Hanselman, from just before the 15-minute mark in:
https://youtu.be/dVG8W-0p6vg?feature=shared&t=896
[*] i try to always use the word "machine"
You see it a LOT in how people talk about natural-selection and evolution in nature. The media, even top rate science media, will talk about evolution as "trying to solve" or "learning" or "adapting" when in fact that is not at all how evolution works. It has no goal, intent, plan, learning, adapting
ChatGPT is an LLM fed by texts, it has no thoughts, only a bunch of uncontrolled texts as input and ditto algorithms designed by uncontrolled commercial entities.
What could possibly go wrong…
I wish I could sit some of these journalists down at chatgpt for 30 minutes and show them how easy it is to lead the model around, producing different kinds of output based on keyword choices.
We are a network of scientists, developers and organizations building the next generation of digital spaces for open science.
