11/09/2024
Welcome to PauseAI DC!
Weโre a grassroots group dedicated to halting the development of the most dangerous AI systems until we know how to make them safe. PauseAI DC is a local branch of PauseAI US, with active groups in cities around the country.
๐๐ก๐ฒ ๐๐ซ๐ ๐ฐ๐ ๐ก๐๐ซ๐?
The world isnโt ready for whatโs coming. A handful of AI companies are racing to build AI systems that can outcompete human beings in all domains. Many experts estimate that human-level AI could only be a few years away.
We do not yet know how to keep advanced AI safe. We can train AI to be intelligent, but we canโt yet reliably control its behavior. Many experts worry that human-level AI could behave unpredictably, taking harmful approaches to achieving its goals. This could mean systems many times smarter than humans, which we cannot control, and which are actively dangerous to us.
We are trying to make something smarter than we are, without knowing how to control it. This is a bad idea.
There are a million more reasons to oppose advanced AI development: the risks of misinformation, deepfakes, and the horrifying prospect of mass job loss, to name a few.
Whatever your biggest concerns, the bottom line is this: we need more time, to set guardrails and to prepare for whatโs next.
--
๐๐ก๐๐ญ ๐๐๐ง ๐ฐ๐ ๐๐จ?
If weโre building something which top scientists think could harm us, the obvious solution is to stop building the thing until we know how to make it safe.
We should pause the development of AI systems past the current frontier, until we have learned how to keep these systems safe and have established guardrails against their misuse.
We also need to work with other countries to set global standards on AI and ensure that no nation develops dangerous AI. Advanced AI knows no borders; it will affect us all, regardless of where it is built.
We have tackled global challenges in the past, such as limiting the spread of nuclear weapons and combating climate change (both are ongoing challenges, of course). We can do it again.
--
๐๐จ๐ฐ ๐๐๐ง ๐ฒ๐จ๐ฎ ๐ก๐๐ฅ๐ฉ?
๐๐๐ฎ๐ฌ๐๐๐ ๐๐ ๐ข๐ฌ ๐ฃ๐ฎ๐ฌ๐ญ ๐ ๐๐ญ๐ญ๐ข๐ง๐ ๐ฌ๐ญ๐๐ซ๐ญ๐๐. ๐๐๐ซ๐ ๐๐ซ๐ ๐ ๐ญ๐ก๐ข๐ง๐ ๐ฌ ๐ฒ๐จ๐ฎ ๐๐๐ง ๐๐จ ๐ญ๐จ ๐ก๐๐ฅ๐ฉ ๐๐๐๐๐:
1. Subscribe to our page for updates. We will have events coming up, including a flyering session and a meeting later this month!
2. Share our page with friends, colleagues, and anyone who may be interested.
3. Share your thoughts in the comments โ what are your concerns around AI, and how would you like to get involved?
Thanks for reading! Stay tuned for much more.
