Trump rejects global AI oversight
3m
President Donald Trump said at the United Nations the United States would call artificial intelligence "superintelligence", rejected a global oversight body and said the Justice Department could intervene.
Keep your expectations low for this week’s visit of XI to the United States when it comes to AI regulation.
But be aware of how high are the stakes for humanity.
Here's what you should know. https://robertreich.substack.com/p/xi-and-the-dilemma-of-ai
But be aware of how high are the stakes for humanity.
Here's what you should know. https://robertreich.substack.com/p/xi-and-the-dilemma-of-ai
September 22, 2026 at 3:48 PM
Keep your expectations low for this week’s visit of XI to the United States when it comes to AI regulation.
But be aware of how high are the stakes for humanity.
Here's what you should know. https://robertreich.substack.com/p/xi-and-the-dilemma-of-ai
But be aware of how high are the stakes for humanity.
Here's what you should know. https://robertreich.substack.com/p/xi-and-the-dilemma-of-ai
Very encouraged to see so many countries coming together to call for mandatory pre-deployment testing & independent evaluation; global coordination on common standards; the creation of an intergovernmental organization; and other critical actions to mitigate the major risks of frontier AI.
A Call for Control of Frontier AI Models - Presidentti
September 2026 Artificial intelligence is an extremely powerful technology. It has potential to improve lives, advance science and strengthen our economies. At the same time, the rapid development of ...
www.presidentti.fi
September 21, 2026 at 8:13 PM
Very encouraged to see so many countries coming together to call for mandatory pre-deployment testing & independent evaluation; global coordination on common standards; the creation of an intergovernmental organization; and other critical actions to mitigate the major risks of frontier AI.
Reposted by Miriam Meckel
Handelsblatt
Creative Destruction: Perma-Frustration and Permadeath – How Could Our Control Over AI Have Gotten So Out of Hand?
The notion that runaway AI agents could suddenly become autonomous is not a law of nature. There are plenty of examples in the history of technology showing how this could be prevented.
Reposts 0
3h
Anyone out there predicting that primary care will be replaced by AI: fine, worried healthy people can talk to bots sometimes, probably.
But have you ever tried managing care or being a caregiver for someone who has multiple conditions and specialists? Like, 25% of Americans?
But have you ever tried managing care or being a caregiver for someone who has multiple conditions and specialists? Like, 25% of Americans?
September 22, 2026 at 5:39 PM
Anyone out there predicting that primary care will be replaced by AI: fine, worried healthy people can talk to bots sometimes, probably.
But have you ever tried managing care or being a caregiver for someone who has multiple conditions and specialists? Like, 25% of Americans?
But have you ever tried managing care or being a caregiver for someone who has multiple conditions and specialists? Like, 25% of Americans?
AI Leaders Are Standing On a Liability Landmine. Today's @wsj.com cites Veronica Paternolli and my paper on liability for agentic AI. www.wsj.com/tech/ai/ai-l...
AI Leaders Are Standing On a Liability Landmine
Plus, Anthropic delays its IPO, and AI agents have trust issues.
www.wsj.com
September 22, 2026 at 5:37 PM
AI Leaders Are Standing On a Liability Landmine. Today's @wsj.com cites Veronica Paternolli and my paper on liability for agentic AI. www.wsj.com/tech/ai/ai-l...
"The proposed channel, which Chinese officials have not publicly agreed to, would allow Bessent and He to contact one another in the event of an artificial intelligence related incident with national security implications."
So not a proper hotline like the Washington-Moscow messaging system, then.
So not a proper hotline like the Washington-Moscow messaging system, then.
‘The scientific, technical term is BS’: Inside the new Bessent-China AI hotline
It is not, at least for now, a broader formal system for the two governments to report or jointly respond to AI threats.
www.politico.com
September 22, 2026 at 5:05 PM
"The proposed channel, which Chinese officials have not publicly agreed to, would allow Bessent and He to contact one another in the event of an artificial intelligence related incident with national security implications."
So not a proper hotline like the Washington-Moscow messaging system, then.
So not a proper hotline like the Washington-Moscow messaging system, then.
Having the AI fight with the insurance companies would be a plus though, no? ;-)
September 22, 2026 at 5:40 PM
Having the AI fight with the insurance companies would be a plus though, no? ;-)
Reposted by Steve Peers
Use of AI to plan violent criminal activities will soon be commonplace. Here's another example: British Columbia sued OpenAI, saying a mass shooting at a school in BC could have been prevented if the company had warned local law enforcement that the shooter had used ChatGPT to plan the massacre.
British Columbia sues OpenAI over Tumbler Ridge school shooting
British Columbia sued OpenAI in California on Monday, saying a mass shooting at a school in the province could have been prevented if the company had warned local law enforcement that the shooter ha...
www.reuters.com
September 22, 2026 at 11:52 AM
Use of AI to plan violent criminal activities will soon be commonplace. Here's another example: British Columbia sued OpenAI, saying a mass shooting at a school in BC could have been prevented if the company had warned local law enforcement that the shooter had used ChatGPT to plan the massacre.
"Any [US-China] A.I. talks will be complicated by the fact that neither country is likely to agree to anything that restrains their own competitive ability, Mr. Mehta said."
Trump and Xi to Discuss A.I. Risks and Competition at US-China Summit (Gift Article)
Artificial intelligence is on the agenda when leaders from the United States and China meet this week. But with little consensus on a path forward, expectations are limited.
www.nytimes.com
September 22, 2026 at 4:04 PM
"Any [US-China] A.I. talks will be complicated by the fact that neither country is likely to agree to anything that restrains their own competitive ability, Mr. Mehta said."
Again, I would genuinely care less about all of this if not for the fact that proponents of weirdo rationalist/EA ideas about how preventing AI from killing us all is The Most Important Thing are now incredibly powerful and wealthy, and will likely get way more so with the Anthropic/OpenAI IPOs.
September 22, 2026 at 5:46 PM
Again, I would genuinely care less about all of this if not for the fact that proponents of weirdo rationalist/EA ideas about how preventing AI from killing us all is The Most Important Thing are now incredibly powerful and wealthy, and will likely get way more so with the Anthropic/OpenAI IPOs.
Reposted by Rosemary A. Joyce
This is a pretty good, succinct explanation of the problem with 2026 effective altruism's pivot from bednets to AI doom risk:
www.washingtonpost.com/opinions/202...
www.washingtonpost.com/opinions/202...
Opinion | This well-meaning ideology fueling AI panic has a dark side
Effective altruism’s prophecy of risks from artificial intelligence could be self-fulfilling.
www.washingtonpost.com
September 22, 2026 at 6:02 PM
This is a pretty good, succinct explanation of the problem with 2026 effective altruism's pivot from bednets to AI doom risk:
www.washingtonpost.com/opinions/202...
www.washingtonpost.com/opinions/202...
101 antitrust initiatives on generative AI by antitrust agencies, Feb 2023 to July 2026.
➝ 2023: United States 7, Europe 7.
➝ 2026 (ongoing): Europe 15, United States 1.
Every enforcement action in the database since November 2025 is European.
➝ 2023: United States 7, Europe 7.
➝ 2026 (ongoing): Europe 15, United States 1.
Every enforcement action in the database since November 2025 is European.
A Database of Antitrust Initiatives Targeting Generative AI - Network Law Review
Antitrust agencies are increasingly interested in generative AI. This can be good news. As Sandy Pentland and I wrote last year (here), the competitive dynamics in this space can be supported by a careful antitrust agenda. While the AI Act should be improved if the EU wants innovation to flourish (see this article), enforcement actions […]
www.networklawreview.org
September 22, 2026 at 12:02 PM
101 antitrust initiatives on generative AI by antitrust agencies, Feb 2023 to July 2026.
➝ 2023: United States 7, Europe 7.
➝ 2026 (ongoing): Europe 15, United States 1.
Every enforcement action in the database since November 2025 is European.
➝ 2023: United States 7, Europe 7.
➝ 2026 (ongoing): Europe 15, United States 1.
Every enforcement action in the database since November 2025 is European.
'Sir Anthony Seldon, former head teacher and author on AI in education, says the technology is being used more and more frequently, but that its introduction has been "chaotic and anarchic".'
Schools 'crying out' for guidance from government on AI, report warns
Author Sir Anthony Seldon says AI needs to be addressed with the same urgency as a war or epidemic.
www.bbc.co.uk
September 22, 2026 at 6:47 AM
'Sir Anthony Seldon, former head teacher and author on AI in education, says the technology is being used more and more frequently, but that its introduction has been "chaotic and anarchic".'
Celone et al. on Architecting Candor
Michael Celone (Massachusetts Institute of Technology (MIT)), Sean McGregor, Mosi Secret, Noga Bregman, Bekhzodkhon Alikhanov, and Eduardo Mignot have posted…...
Michael Celone (Massachusetts Institute of Technology (MIT)), Sean McGregor, Mosi Secret, Noga Bregman, Bekhzodkhon Alikhanov, and Eduardo Mignot have posted…...
Celone et al. on Architecting Candor
Michael Celone (Massachusetts Institute of Technology (MIT)), Sean McGregor, Mosi Secret, Noga Bregman, Bekhzodkhon Alikhanov, and Eduardo Mignot have posted Architecting Candor: Products Liability and AI Incident Knowledge Governance on SSRN. Here is the abstract: Artificial intelligence (AI) firms need incident knowledge to improve safety, yet the act of documenting that knowledge can increase litigation risk. As American courts increasingly subject AI systems to products liability, and as the European Union (EU)’s Product Liability Directive expressly classifies software as a “product,” the root-cause analysis necessary to diagnose an incident and remediate the system provides plaintiffs with evidence to establish fault and defective design.
legaltheoryblog.com
September 21, 2026 at 11:05 PM
Celone et al. on Architecting Candor
Michael Celone (Massachusetts Institute of Technology (MIT)), Sean McGregor, Mosi Secret, Noga Bregman, Bekhzodkhon Alikhanov, and Eduardo Mignot have posted…...
Michael Celone (Massachusetts Institute of Technology (MIT)), Sean McGregor, Mosi Secret, Noga Bregman, Bekhzodkhon Alikhanov, and Eduardo Mignot have posted…...
With $7 trillion on the line, AI safety may always take a backseat to market domination www.axios.com/2026/09/22/a...
With $7 trillion on the line, AI safety may always take a backseat to market domination
Why slowing down AI scaling may be impossible.
www.axios.com
September 22, 2026 at 10:43 AM
With $7 trillion on the line, AI safety may always take a backseat to market domination www.axios.com/2026/09/22/a...
Reposted by Iikka Korhonen
EXCLUSIVE: Scott Bessent proposed a new AI safety “notification mechanism” during a Sunday meeting with Chinese Vice Premier He Lifeng.
But the proposal is far simpler than the term suggests: an open line of communication between the two officials.
But the proposal is far simpler than the term suggests: an open line of communication between the two officials.
‘The scientific, technical term is BS’: Inside the new Bessent-China AI hotline
It is not, at least for now, a broader formal system for the two governments to report or jointly respond to AI threats.
www.politico.com
September 22, 2026 at 4:25 PM
EXCLUSIVE: Scott Bessent proposed a new AI safety “notification mechanism” during a Sunday meeting with Chinese Vice Premier He Lifeng.
But the proposal is far simpler than the term suggests: an open line of communication between the two officials.
But the proposal is far simpler than the term suggests: an open line of communication between the two officials.
Reposted by Hilary J. Allen
From @timnitgebru.blacksky.app and me on what we should take from the 2026 summer of AI hype:
www.technologyreview.com/2026/09/22/1...
www.technologyreview.com/2026/09/22/1...
Don’t be fooled by this summer of AI hype
Breathless claims about AGI and new capabilities fall apart pretty quickly under scrutiny.
www.technologyreview.com
September 22, 2026 at 1:00 PM
From @timnitgebru.blacksky.app and me on what we should take from the 2026 summer of AI hype:
www.technologyreview.com/2026/09/22/1...
www.technologyreview.com/2026/09/22/1...
Reposted by Lesley A. Hall
Watching doctors on the BBC saying how dangerous & harmful AI is proving by giving vulnerable people false medical advice. They say it's bad, but AI is here & we must adapt to it. It's as if governments, laws, & regulations don't exist - we've been convinced by tech-bros to be helpless.
September 22, 2026 at 7:27 AM
Watching doctors on the BBC saying how dangerous & harmful AI is proving by giving vulnerable people false medical advice. They say it's bad, but AI is here & we must adapt to it. It's as if governments, laws, & regulations don't exist - we've been convinced by tech-bros to be helpless.
Looks like his job can be replaced by AI.
September 21, 2026 at 8:30 PM
Looks like his job can be replaced by AI.
Reposted by Steven H. Seggie
Staff at world-leading AI companies and institutions are increasingly suffering from the mental strain of working on systems they fear could cause serious harm to society
www.ft.com/content/6087...
www.ft.com/content/6087...
AI staff complain of mental toll over fears of threat to society
Top researchers at Aisi, OpenAI, Anthropic and Google DeepMind report burnout and stress over developing powerful AI
www.ft.com
September 22, 2026 at 6:39 AM
Staff at world-leading AI companies and institutions are increasingly suffering from the mental strain of working on systems they fear could cause serious harm to society
www.ft.com/content/6087...
www.ft.com/content/6087...