PauseAI London Protest
The next AI safety summit is held in San Francisco this November, so we need to remind our leaders of their most important responsibility. We ask the government to "Come back from San Francisco with a treaty". Join us on November 21st to get this message across to them.
AI companies are locked in a race to develop superintelligence as fast as possible, despite the fact that most researchers think that AI might cause human extinction (https://blog.aiimpacts.org/p/2023-ai-survey-of-2778-six-things). This is a crazy situation! But we can do something to help.
1 - **We don't understand how AI works.** Currently AIs are black boxes - we know how to train them but we cannot interpret the internal algorithms and representations of frontier systems.
2 - **We don't have rigorous guarantees that AI can be controlled.** We need a detailed science of AI control, not vague fine-tuning.
3 - **We do not have governance mechanisms to ensure that AI will be deployed for everyone's benefit.** AI has the power to solve many problems and relieve suffering. But it could also cause mass unemployment, spread misinformation, be used in autonomous weapons and concentrate power into the hands of a few companies.
We ask the government to come back from San Francisco with a treaty pausing frontier AI development until proof of safety. Read our full proposal here: https://pauseai.info/proposal
Schedule:
4:30pm - Main event
5:30pm - Pub social - The Red Lion, Parliament Street