Tag: Maggie Harrison Dupré

Robot Therapist AI Mental Healthcare Concept Future Psychology Support Empathy Technology Listening Session

^{Type
post

News
Date
August 20, 2025

Categorized
Artificial general intelligence (AGI), Business and Finance, Large Language Models (LLMs), Psychology, Relationships

Tagged
AI and values, AI bubble, AI therapists, Chris Rufo, Gary Marcus, Geoffrey Hinton, J. David Creswell, Maggie Harrison Dupré, My Boyfriend Is AI (Reddit), Sam Altman}

AI Is Impacting Our Culture — But Not As We Were Told To Expect

_{Okay, your boss is not a robot. Your doctor is not a chatbot. But here are some things that are actually happening that we might not have expected} _{News
August 20, 2025

5

Artificial general intelligence (AGI), Business and Finance, Large Language Models (LLMs), Psychology, Relationships}

One change is that Sam Altman, CEO of OpenAI, the maker of ChatGPT-5, seems to be acknowledging that the AI bubble might soon burst. Read More ›

^{Type
post

News
Date
August 28, 2024

Categorized
Creativity, Large Language Models (LLMs)

Tagged
Maggie Harrison Dupré, Model collapse}

AI Going MAD? The Model Collapse Problem Gets More Attention Now

_{The problem is that the AI can't generate anything new on its own and the old stuff, endlessly recycled at high speeds, begins to erode} _{News
August 28, 2024

3

Creativity, Large Language Models (LLMs)}

If human involvement is minimized, this hard limit will probably come up time and again. Computation is not creativity. Read More ›

cute artificial intelligence robot with notebook

^{Type
post

News
Date
June 13, 2024

Categorized
Artificial Intelligence

Tagged
Chatbots and logic puzzles, Maggie Harrison Dupré}

Researchers: Chatbots Generally Flunk Simple Logic Test

_{From Futurism we learn “when asked to show their work, the AIs would detail bizarre and erroneous lines of "thinking" that made no sense”} _{News
June 13, 2024

4

Artificial Intelligence}

The AI researchers report “a dramatic breakdown of function and reasoning capabilities of state-of-the-art models trained at the largest available scales”. Read More ›