Grassroots movement to pause dangerous AI development until it's safe. Organizing local groups and connecting with our representatives across the US.
United States
Joined January 2024
- Tweets923
- Following194
- Followers1.8K
- Likes1.9K
Pinned Tweet
Today we held a demonstration at the State Department in Washington D.C. in support of a U.S.-led international treaty to pause AI.
Things are changing fast. The public is realizing the enormous danger of building Artificial Intelligence. Congress is preparing to act. The American people are on our side. There is much reason for hope.
We're proud of our volunteers who showed up, and of everyone on earth who will help humanity avert an AI catastrophe and pause AI now.
PauseAI US ⏸️ retweeted
In redder-than-red Idaho, this is a great sign that people are concerned and don't see AI risks primarily as a partisan issue. Well done everyone!
Proud of our volunteers in Boise who organized Idaho's first AI Safety protest last weekend! PauseAI NOW!
The world is listening, and our government is listening to our concerns about the dangers of AI. Anyone can help, and if you want to help, join us this December as we go to meet with our elected representatives in your home district!
We will train you, book your meetings, and connect you with other local volunteers. Apply for PauseCon Home here: pauseai-us.org/pausecon
PauseAI US ⏸️ retweeted
Proud of our Boise volunteers!! Join us, meet directly with Congress in your home district. They are listening, and you can help pass AI regulation. Apply here: pauseai-us.org/pausecon
Proud of our volunteers in Boise who organized Idaho's first AI Safety protest last weekend! PauseAI NOW!
The world is listening, and our government is listening to our concerns about the dangers of AI. Anyone can help, and if you want to help, join us this December as we go to meet with our elected representatives in your home district!
We will train you, book your meetings, and connect you with other local volunteers. Apply for PauseCon Home here: pauseai-us.org/pausecon
PauseAI US ⏸️ retweeted
So beyond proud of our Boise volunteers, especially our stunning husband-wife organizing team Jack and Cathryn 🥹
Proud of our volunteers in Boise who organized Idaho's first AI Safety protest last weekend! PauseAI NOW!
The world is listening, and our government is listening to our concerns about the dangers of AI. Anyone can help, and if you want to help, join us this December as we go to meet with our elected representatives in your home district!
We will train you, book your meetings, and connect you with other local volunteers. Apply for PauseCon Home here: pauseai-us.org/pausecon
Proud of our volunteers in Boise who organized Idaho's first AI Safety protest last weekend! PauseAI NOW!
The world is listening, and our government is listening to our concerns about the dangers of AI. Anyone can help, and if you want to help, join us this December as we go to meet with our elected representatives in your home district!
We will train you, book your meetings, and connect you with other local volunteers. Apply for PauseCon Home here: pauseai-us.org/pausecon
The AI companies are "wargaming scenarios for the aftermath of a catastrophic event." We need to PauseAI NOW! Before any catastrophe. Pause saves lives.
Join us in Boise, October 24th at 11am, for Idaho's first AI Safety protest. Organized by our amazing Local Group Leaders.
Learn more: pauseai-us.org/protest
PauseAI US ⏸️ retweeted
OpenAI ‘agents’ disabled their own monitoring systems ft.trib.al/jzXsiSj
If you're in Boise this weekend, come to our protest and make your voice heard! pauseai-us.org/protest
For everyone else in the US, apply for PauseCon! Four times a year we organize to meet directly with our elected representatives. We're up to 294 meetings with government! Apply here: pauseai-us.org/pausecon
Sharing more pictures of our wonderful volunteers, who traveled to Washington DC last month and met with 42 Congressional Offices from 15 different states!
Our protest in Boise, Idaho is THIS SATURDAY! We've organized demonstrations in San Francisco, Washington DC, Phoenix, Providence, Anchorage, and now Boise! Come out and make your voice heard, tell our government to PauseAI NOW! pauseai-us.org/protest
The AI companies are "wargaming scenarios for the aftermath of a catastrophic event." We need to PauseAI NOW! Before any catastrophe. Pause saves lives.
Join us in Boise, October 24th at 11am, for Idaho's first AI Safety protest. Organized by our amazing Local Group Leaders.
Learn more: pauseai-us.org/protest
Thank you to @donatedotgg for sending these funds to us! This has become our largest fundraiser of the year. We will use these funds to further our mission of pausing AI, through advocacy and local organizing!
PauseAI US ⏸️ retweeted
So, any former Anthropic stans want to apologize? You really did a bad thing supporting them all this time bc it protected them from resistance. It was not costless or safe to take the position of trusting Anthropic and a good person who realizes this will feel remorse.
PauseAI US ⏸️ retweeted
Mechanistic interpretability is not going to make anything safer if people are still able to apply their goals of taking dangerous risks to the AI system. Knowing more about AI is fundamentally dual use and could make it MORE dangerous. What makes AI safe (or keeps it paused forever, if necessary) is proper governance to keep it in line with the interests of the people it affects.
There was a time when it seemed like the problem would be the alienness of the AI itself. But even if AI were fully predictable and under our control we have no solved the problem of how some people want to use that power. Interpretability is not just a knowledge for safety— it could be the biggest capabilities unlock we’ve ever seen. The AI companies getting it is not something to celebrate.
PauseAI US ⏸️ retweeted
Yeah I don’t really trust @OpenAI saying “🤩 get ready to stop worrying and trust OpenAI again 🤩”
PauseAI US ⏸️ retweeted
A major project during a slowdown (a pause) is transforming the chip supply chain from a general-purpose chip that can do training and inference, to model-specific chips that can only do inference for a specific model. That way we can use existing capabilities, but can’t seamlessly keep escalating intelligence. This would also prevent reckless U.S.-China competition.
You can read more about this proposal co-led by Berkley professors @wfithian and @ProfHolliday here: hardwired-pause.ai
