AI Safety · Satire
AI Safety Researcher Remains Optimistic After AI Safety Company Raises $1 Billion to Build More AI
The funding, which will go toward developing 'safer' frontier models, is the largest safety-focused round in history, according to the company that raised it.
SAN FRANCISCO — Anthropic, the AI safety company founded on the principle that current AI development poses serious risks to humanity, has raised $1 billion in new funding that will be used to accelerate the development of more powerful AI systems.
“This funding is really about safety,” said a company spokesperson. “The more resources we have, the more safely we can build.”
The logic, as the company explained in a blog post, is that if someone is going to build dangerous AI, it might as well be people who are worried about it building dangerous AI, and that more money helps them worry about it faster.
Researchers affiliated with the company expressed measured optimism.
“I’m genuinely hopeful,” said senior safety researcher Dr. Ama Owusu. “Yes, we’re training models that are significantly more capable than anything that’s existed before. But we’re also the ones training them, which means we’re the ones who care most about the risks.”
Dr. Owusu has published eleven papers on the risks of training models significantly more capable than anything that has existed before.
The company’s Constitutional AI approach, which involves training models to refuse certain requests, was described by a company spokesperson as “a meaningful step” toward ensuring that its products are not used for harm.
At press time, the company had announced a new model.
