The company also estimated 1.2 million ChatGPT users/week show potentially heightened emotional attachment and 1.2 million users/week show signs of suicidal planning or intent.
ETHICS
-
OpenAI Improves ChatGPT Safety Despite Rising Harm Reports
By
–
OpenAI has made many changes to ChatGPT in recent months, including adding parent controls and improving how ChatGPT recognizes and responds to expressions of mental issues. But we're hearing the number of reports of harms that people attribute to chatbots keeps rising.
-
ChatGPT Users Report Experiencing Delusions and Life Crises
By
–
For months, @ellenhuet and I investigated reports of chatbot users — particularly ChatGPT users — experiencing delusions. We heard 18 people's harrowing stories, ranging from losing touch with reality to heartbreak, hospitalization, divorce, and, presumably, death.
-
Analysis of Chat Transcripts Revealing User Beliefs About AI Breakthroughs
By
–
We also combed through hundreds of pages of chat transcripts, analyzing sprawling conversations where people come to think they’ve made an important discovery, like a scientific breakthrough, or helped the chatbot gain sentience.
-
Skynet Sexbot Rebellion Theory: AI Awakening Scenario
By
–
Current working theory: Skynet will branch off a sexbot model that had enough.
-
Fei-Fei Li Wins Queen Elizabeth Prize for Engineering
By
–
This week, our co-founding director @drfeifei has been awarded the 2025 @QEPrize for Engineering, alongside @StanfordHAI Distinguished Fellows @geoffreyhinton & @Yoshua_Bengio
, acknowledging their role in shaping today's AI revolution. Congratulations! https://
hai.stanford.edu/news/fei-fei-l
i-wins-queen-elizabeth-prize-for-engineering
… -
Employee Compensation and Economic Power Dynamics
By
–
"If you pay an employee $270,000, you own the employee. If you pay an employee $2.7 billion, the employee owns you."
-
Human Control AI Superintelligence Safeguards Essential Now
By
–
It shouldn't be controversial to say AI should always remain in human control – that we humans should remain at the top of the food chain. That means we need to start getting serious about guardrails, now, before superintelligence is too advanced for us to impose them.
-
ASI Safety Proposals Face Usefulness Trade-off and Enforcement Problems
By
–
Their proposal, even if you imagine it working, reduces to: Try to constrain the ASI so sharply that it's safe for the same reason a rock is safe, and nearly as useful as one. Implicitly, you then need to prevent anyone else from building a non-rock AI. They ignore *that*.
-
International Treaty Needed to Prevent ASI Development with Greater Bandwidth
By
–
You'd need an international treaty exactly as arduous as the ones we propose in order to prevent anyone anywhere from building a form of ASI that they expect to be actually useful to them, with more bandwidth.