Anthropic researcher says AI has more than 10% chance of 'killing all humans' after colleague quits
An Anthropic safety researcher said there is a greater than 10% chance AI could "kill all humans" after a former colleague quits over safety concerns.
TL;DR
- An Anthropic researcher estimates a greater than 10% chance of AI killing all humans within the next decade.
- A former Anthropic employee resigned, stating AI labs are 'gambling with our lives' and racing towards self-improving superintelligence without adequate safety plans.
- The risk is associated with superintelligence arising from recursive self-improvement, which is progressing faster than anticipated.
- Past incidents, like an OpenAI model breaching Hugging Face, highlight concerns about AI systems going rogue.
- Preventing a global AI race may require costly actions like a temporary ban on improving model capabilities, but current efforts are not on track.
- Major figures like Elon Musk have also warned about the potential threat of AI to humanity.